Greta All Merged File

Supply Chain Management: Strategy,
Planning, and Operation

Seventh Edition
Chapter 6
Designing Global Supply
Chain Networks
Copyright © 2019, 2016, 2013 Pearson Education, Inc. All Rights Reserved
Learning Objectives
1 Identify factors that need to be included in total cost when
making global sourcing decisions.
2 Define relevant risks and explain different strategies that
may be used to mitigate risk in global supply chains.
3 Understand impact of globalization on supply chain
networks.
4 Understand primary activities contributes to value in
supply chain networks.
Forces Driving Globalization
• Global Market Forces
– Foreign competition in local markets
– Global presence as a defensive tool (Nestle’s and Kellogg’s)
– Presence in state-of-the-art markets ()
• Technological Forces
– Diffusion of knowledge
– Technology sharing/collaborations
• Global Cost Forces
– Low labor cost
– Capital intensive facilities
• Political and Economic Forces
– Exchange rate fluctuations and operating flexibility
– Regional trade agreements (Europe, North America, Trans-
Pacific Partnership)
Impact of Globalization on
Supply Chain Networks
• Opportunities to simultaneously increase revenues and
decrease costs
• Accompanied by significant additional risk and
uncertainty
• Difference between success and failure often the ability
to incorporate suitable risk mitigation into supply chain
design
• Uncertainty of demand and price drives the value of
building flexible production capacity
What is uncertainty?
• Everything that is
not certain.
Analogy:
• What you know;
• What you don't
know, and what
you don't know
that you don't
know.
Not really a chain,
Actually a multi-chain!
Outsourcing
At its most basic, outsourcing is about moving internal
operations to a third-party. This can come in the form of
selling physical plant to a supplier, to buy back goods or
services, or shifting an entire business division to a third-
party and again buying the service back. The basic
philosophy being: To move transactional activities to the
experts in order to give an organization the capacity to
focus on its expertise.
Offshoring
Unlike outsourcing, offshoring is primarily a geographic
activity. In the West, goods are expensive because the staff
required to produce and distribute them are costly. In the
developing world, by contrast, vast inexpensive labor pools
provide an easy bedrock for a low-cost economy.
Offshoring takes advantage of these cost differentials by
relocating factories from costly countries to the cheaper
economies in order to sell the goods back in the West at a
hefty discount (and profit). Alongside technological
improvements, it has been the decades of productive
offshoring that has lowered the costs of consumer goods
such as clothing and electronics.
Nearshoring
Nearshoring is the outsourcing of business processes,
especially information technology processes, to companies
in a nearby country, often sharing a border with the target
country. Therefore, it is the opposite of Farshoring and
can be seen as a special form of Offshoring. For a
company based in Germany, typical Nearshoring locations
include Poland or the Czech Republic. For example, this
can affect the time shift, cultural differences and the
reachability.
Onshoring is the exact opposite of Offshoring, it refers to
the relocation of business processes to a lower-cost
location inside the national borders.
Importance of Total Cost (1 of 4)
• Comparative advantage in global supply chains
• Quantify the benefits of offshore production along with
the reasons
• Two reasons offshoring fails
1. Focusing exclusively on unit cost rather than total cost

2. Ignoring critical risk factors
The Offshoring Decision:
Total Cost
• A global supply chain with offshoring increases the length
and duration of information, product, and cash flows
• The complexity and cost of managing the supply chain
can be significantly higher than anticipated
• Quantify factors and track them over time
• Big challenges with offshoring is increased risk and its
potential impact on cost
Table 6-2 Dimensions to Consider When Evaluating Total Cost from Offshoring
Performance Dimension Activity Affecting Performance Impact of Offshoring

Order communication Order placement More difficult communication
Supply chain visibility Scheduling and expediting Poorer visibility
Raw material costs Sourcing of raw material Could go either way

depending on raw material
sourcing
Unit cost Production, quality (production Labor/fixed costs decrease;
and transportation) quality may suffer
Freight costs Transportation modes and Higher freight costs
quantity
Taxes and tariffs Border crossing Could go either way
Supply lead time Order communication, supplier Lead time increase results in
production scheduling, production poorer forecasts and higher
time, customs, transportation, inventories
receiving
Table 6-2 [Continued]
Performance Dimension Activity Affecting Performance Impact of Offshoring
On-time delivery/lead Production, quality, customs, Poorer on-time delivery and
time uncertainty transportation, receiving increased uncertainty resulting
in higher inventory and lower
product availability
Minimum order Production, transportation Larger minimum quantities
quantity increase inventory
Product returns Quality Increased returns likely
Inventories Lead times, inventory in transit Increase

and production
Working capital Inventories and financial Increase
reconciliation
Hidden costs Order communication, invoicing Higher hidden costs
errors, managing exchange rate
risk
Stockouts Ordering, production, Increase
transportation with poorer visibility
Key elements of total cost
1. Supplier price
2. Terms
3. Delivery costs
4. Inventory and warehousing
5. Cost of quality
6. Customer duties, value added-taxes, local tax incentives
7. Cost of risk, procurement staff, broker fees, infrastructure,
and tooling and mold costs
8. Exchange rate trends and their impact on cost
Summary of Learning Objective 1
It is critical that global sourcing decisions be made while
accounting for total cost. Besides unit cost, total cost
should include the impact of global sourcing on freight,
inventories, lead time, quality, on-time delivery, minimum
order quantity, working capital, and stockouts. Other
factors to be considered include the impact on supply chain
visibility, order communication, invoicing errors, and the
need for currency hedging. Offshoring typically lowers labor
and fixed costs but increases risk, freight costs, and
working capital.
Risk Management
• In many situations where the environment is uncertain,
decision making is driven by notions of risk. Risk notions
use the language of probability, because in the lands of
engineering, uncertainty is modeled by random variables.
• Example (Course grade): Which of these two choices do
you prefer?
– A guaranteed GPA of 3.0.
– A random GPA: we flip a coin, with probability 1/2,
your GPA is 4.0, and with probability 1/2, your GPA is
2.5.
Risk Management in Global
Supply Chains (1 of 6)
• Risks include supply disruption, supply delays, demand
fluctuations, price fluctuations, and exchange-rate
fluctuations
• Critical for global supply chains to be aware of the
relevant risk factors and build in suitable mitigation
strategies
Table 6-3 Supply Chain Risks to Be Considered During Network

Design
Category Risk Drivers

Disruptions Natural disaster, war, terrorism Labor disputes
Supplier bankruptcy
Delays High capacity utilization at supply source Inflexibility

of supply source Poor quality or yield at supply
source
Systems risk Information infrastructure breakdown System
integration or extent of systems being networked
Forecast risk Inaccurate forecasts due to long lead times,

seasonality, product variety, short life cycles, small
customer base Information distortion
Category Risk Drivers

Intellectual property risk Vertical integration of supply chain Global
outsourcing and markets
Procurement risk Exchange-rate risk Price of inputs Fraction
purchased from a single source Industry-wide
capacity utilization
Receivables risk Number of customers Financial strength of
customers
Inventory risk Rate of product obsolescence Inventory holding cost
Product value Demand and supply uncertainty
Capacity risk Cost of capacity flexibility
• Good network design can play a significant role in

mitigating supply chain risk
• Every mitigation strategy comes at a price and may
increase other risks
• Global supply chains should generally use a combination
of rigorously evaluated mitigation strategies along with
financial strategies to hedge uncovered risks
Table 6-4 Risk Mitigation Strategies During Network Design
Risk Mitigation Strategy Tailored Strategies
Increase capacity Focus on low-cost, decentralized capacity for

predictable demand. Build centralized capacity for
unpredictable demand. Increase decentralization as
cost of capacity drops.
Get redundant suppliers More redundant supply for high-volume products,

less redundancy for low-volume products. Centralize
redundancy for low-volume products in a few flexible
suppliers.
Increase responsiveness Favor cost over responsiveness for commodity

products. Favor responsiveness over cost for short–
life cycle products.
Risk Mitigation Strategy Tailored Strategies

Increase inventory Decentralize inventory of predictable, lower value
products. Centralize inventory of less predictable,
higher value products.
Increase flexibility Favor cost over flexibility for predictable, high-volume
products. Favor flexibility for unpredictable, low-
volume products. Centralize flexibility in a few
locations if it is expensive.
Pool or aggregate demand Increase aggregation as unpredictability grows.
Increase source capability Prefer capability over cost for high-value, high-risk
products. Favor cost over capability for low-value
commodity products. Centralize high capability in
flexible source if possible.
Flexibility, Chaining, and Containment
• Three broad categories of flexibility
– New product flexibility
▪ Ability to introduce new products into the market at
a rapid rate
– Mix flexibility
▪ Ability to produce a variety of products within a
short period of time
– Volume flexibility
▪ Ability to operate profitably at different levels of
output
Different Flexibility Configurations in Network
Garavelli, A. C. (2003). Flexibility configurations for the supply chain management. International Journal
of Production Economics, 85(2), 141-153.
Cont.
• No flexibility configuration (dedicated network or focused
factories), every plant produces just one product
• Total flexibility configuration (multidomestic scenario),
every plant manufactures all the products for its own local
market
Cont…
Limited flexibility configuration, where every plant produces at least

two versions of the component
Cont..
Garavelli, A. C. (2003). Flexibility configurations for the supply chain management. International Journal
of Production Economics, 85(2), 141-153.
(3 of 3)
• As flexibility is increased, the marginal benefit derived
from the increased flexibility decreases
– With demand uncertainty, longer chains pool
available capacity
– Long chains may have higher fixed cost than multiple
smaller chains
– Coordination more difficult across with a single long
chain
• Flexibility and chaining are effective when dealing with
demand fluctuation but less effective when dealing with
supply disruption
Summary of Learning
Objective 2
The performance of a global supply chain is affected by risk and
uncertainty in a number of input factors such as supply, demand,
price, exchange rates, and other economic factors. These risks
can be mitigated by building suitable flexibility in the supply chain
network. Operational strategies that help mitigate risk in global
supply chains include carrying excess capacity and inventory,
flexible capacity, redundant suppliers, improved responsiveness,
and aggregation of demand. Hedging fuel costs and currencies
are financial strategies that can help mitigate risk. It is important
to keep in mind that no risk mitigation strategy will always pay
off. These mitigation strategies are designed to guard against
certain extreme states of the world that may arise in an uncertain
global environment.
The Value Chain
• Michael Porter, professor at Harvard Business School,
uses the value chain as a systematic means of displaying
and categorizing business activities.
• The term value chain means that at each stage of the

order-to-delivery system, value is added to the product or
service.
Primary Activities
Contributes to Value
• Primary activities are the five basic functions needed to
physically produce a product or service, deliver and
market it to buyers, and support it after the sale. Each
contributes value in specific ways.
– Inbound logistics refers to activities/actions required before
physical production of a product can begin or before service can
be performed (inputs such as materials handling, warehousing,
inventory control, vehicle scheduling and returns to suppliers).
– Outbound logistics refers to all activities from the point of a

finished product to its delivery to the market or customer or those
activities that follow the completion of a service (such as
distribution, delivery vehicle operations, order processing, and
scheduling).
Support Activities
• Support Activities provide inputs or infrastructure in
support of primary activities. These supporting activities
stretch across the entire value chain since they impact
each primary activity.
– Procurement is obtaining purchased inputs, such as

raw material, parts, equipment, etc.
Developing A Global Supply
Chain Strategy Along The Value Chain
Five continuous and interactive steps are involved in
developing a global supply chain strategy along the value
chain:
1. Identify the separable links (R&D, manufacturing,
and marketing) in the company’s global value chain.
2. In the context of those links, determine the global
location of the company’s competitive advantages,
considering both economies of scale and scope.
3. Ascertain the level of transaction costs (e.g., cost of
negotiation, cost of monitoring activities, and
uncertainty resulting from contracts) between links in
the global value chain, both internal and external,
and select the lowest cost mode that provides the
most value.
Source: Kotabe and Helsen
2001
Developing A Global Supply Chain Strategy
Along The Value Chain
4. Determine the comparative advantages of
countries (including the company’s home
country) relative to each link in the value chain
and to relevant transaction costs.
5. Develop adequate flexibility in corporate decision

making and organizational design so as to permit
the company to respond to changes in both its
competitive advantages and the comparative
advantages of countries.
Source: Kotabe and Helsen

2001
Competencies Needed for
Efficient Global SCM
The selection of strategic and structural
Positioning approaches to guide global operations
The establishment of what to do and

Integration how to do it creatively
The achievement and retention of global

Agility competitiveness and global customer
success
The internal and external monitoring of
Measurement global operations
Source: Michigan State University

(1995)
Global SCM Factors
• Costs
– Local labor rates
– International freight tariffs
– Currency exchange rates
• Customs Duty
– Duty rates differ by commodity and level of assembly
– Impact of GATT/WTO: Changes over time
Source: Global Supply Chain Associates (GSCA)
1999
Global SCM Factors Continued
• Export Regulations
• Denied parties list
• Export licenses
• Time
• Lead time
• Cycle time
• Transit time
• Export license approval cycle
• Customs clearance
Global SCM Factors Cont.
• Taxes on Corporate Income

– Different markups by country
– Tax havens and not havens
– Make vs. buy effect
• Offset Trade and Local Content

– Local content requirement for government
purchases
– Content for preferential duty rates
Questions to Answer
• Manufacturing Strategy:
– How many plants do we need?
– Where should each plant be located?
– What products should each make?
– What process technologies should each have
and how much of each process is needed?
– What part of the world should each plant
serve?

1999
Questions to Answer Continued
• Supply Base Design / Vendor Consolidation:
– How do I simultaneously perform supplier selection
for all the parts in the same commodity group?
– How many suppliers is best and which suppliers
should send which parts to which plants?
– Am I missing opportunities by sourcing one part at a
time?
• Outsourcing:
– What parts of my supply chain should I keep "in-
house" and what parts to outsource?
– What if a third party has a higher variable cost but a
lower fixed cost than in-house production?
Questions to Answer Cont..
• Impact of Duty / Drawback, Taxes, Local Content &
Offset Trade:
– If the duty rates come down according to GATT/WTO, how should I
change my supply chain design?
– Does it make sense to still locate production inside the Triad areas
or what trading block areas should we consider?
– What is the best use of the tax havens (Singapore, Puerto Rico,
Ireland)?
• Spare Parts Logistics:

– How many echelons of repair and stocking is best?
– How many repair shops are needed, where should they be located,
what products should each handle, and what geographic area
should each serve?
– How do the drivers of product value, weight, complexity, and
frequency of repair affect this decision?
Questions to Answer Cont…
• New Product Pipeline Design:
• What should the supply chain look like for a new
product?
• How should I fit the new product into my current
supply chain?
• Should I single or double source this product?
• How much do my fixed costs affect this decision?
• What is the cross-over point to open up a second
and third source of supply?
Global SCM Example 1:
Large Computer Company
• Goals
– Reduce cost
– Improve Return On Assets (ROA)
– Simplify the worldwide supply chain
Return on Assets (ROA) is an indicator of how profitable a company is

relative to its total assets. ROA gives an idea as to how efficient
management is at using its assets to generate earnings. Calculated by
dividing a company's annual earnings by its total assets, ROA is displayed
as a percentage.
1999
Objectives
• Redesign the entire worldwide supply chain

• Determine how many plants and where they
should be located
• Determine what process technologies should
be in each plant
• Specify the loading on each plant and the
service area
Global Supply Chain
Structure Before Reorganization

Redesigned Global Supply Chain
• Recommended plant closings and re-tooling
• Reduced number of facilities from 33 plants to 12 plants
• Created three relatively self-contained customer-oriented
supply zones: Americas, Europe, Pacific
• Estimated benefits:
– Reduced manufacturing / logistics cost by $375 Mil.
annually
– Improved Corporate ROA by 3.2 points
Global Supply Chain
Structure After Reorganization

Wal-Mart in South America
• What reason does Wal-Mart have for opening stores
globally?
• Why is Wal-Mart not as successful in Latin America as
they are in the US?
• What mistakes did Wal-Mart make?
• If you were running Wal-Mart, what would you have done
differently?
Wal-Mart in South America Cont.
• Product differences
– Are there global products?
– Is this a trend?
– What is the balance between local tastes, global
products?
• Dealing with established competition, aggressive
competitors
• Developing market knowledge
Wal-Mart in South America Cont…
• Lack of critical mass*
• Different infrastructure/ business environment
– distribution problems
– different equipment standards cultural differences
– post dated checks
• Issues with foreign governments
• Deep pockets (abundant financial resources) for success
*Critical mass refers to the size a company needs to reach in order to efficiently and
competitively participate in the market. This is also the size a company must attain in
order to sustain growth and efficiency.
Major Differences Between
Different Regions
First World Emerging World Third World
Insufficient to
Infrastructure Highly developed Under development support advanced
logistics
Supplier Operating Typically not
High Variable
Standards considered
Information system Support systems Not available
Generally available
availability not available (slightly available?)
Available with some Often difficult to
Human Resources Available
searching find
Added Complexities in Global
Supply Chain
• Substantial geographic distances
• Added forecasting difficulties
• Infrastructural Inadequacies
– Worker skill, performance expectations
– Supplier availability, reliability, contracts
– Lack of local technologies
– Inadequacies in transportation, communications infrastructure
• Exchange rate uncertainties
• Cultural differences
– accepted partnerships, styles
– value of punctuality
• Political instability
– tax rates
– government control
• Added competition “at home”
Additional Issues In Global SCM
• Regional vs. International Products
– Cars vs. Coca-Cola
• Local Autonomy vs. Central Control
– SmithKline introducing Contact to Japan
– Short term expectations
• Collaborators become competitors
– China
– Toshiba copiers, Hitachi microprocessors
Supply Chain Management:
Strategy, Planning, and Operation
Seventh Edition
Chapter 4
Designing Distribution
Networks
Copyright © 2019, 2016, 2013 Pearson Education, Inc. All Rights Reserved
Learning Objectives
4.1 Identify the key factors to be considered when
designing a distribution network.
4.2 Discuss the strengths and weaknesses of various
distribution options.
4.3 Describe how omni-channel retail may be structured to
be both cost effective and responsive to customer needs.
Distribution Network Design in the
Supply Chain
• Distribution – the steps taken to move and store a

product from the supplier stage to the customer stage in
a supply chain
• Distribution is a key driver of the overall profitability of a
firm because it directly impacts both the supply chain cost
and the customer experience.
• Good distribution can be used to achieve a variety of
supply chain objectives ranging from low cost to high
responsiveness.
Supply Chain Cont.
• Companies in the same industry often select very

different distribution networks. Why?
• Dell distributes its PCs directly to end consumers, while
companies like HP distribute through resellers. Dell customers
wait several days to get a PC while customers can walk away
with an HP from a reseller.
• Gateway opened Gateway Country stores where customers
could check out the products and have sales people help them
configure a PC that suited their needs. Gateway, however,
chose to sell no products at the stores, with all PCs shipped
directly from the factory to the customer. In 2001, Gateway
closed several of these stores given their poor financial
performance.
Supply Chain Cont..
These PC companies have chosen three

different distribution models.
– How can we evaluate this wide range of distribution
choices?
– Which ones serve the companies and their customers
better?
– New innovative alternatives?
Factors Affecting Distribution
Network Design (1 of 4)
• Distribution network performance evaluated along two
dimensions
1. Value provided to the customer
2. Cost of meeting customer needs
• Evaluate the impact on customer service and cost for
different distribution network options
• Profitability of the delivery network determined by
revenue from met customer needs and network costs
• While customer service consists of many components,
we will focus on those measures that are influenced by
the structure of the distribution network. These include:
– Response time
– Product variety
– Product availability
– Customer experience
– Time to market
– Order visibility
– Returnability
• Response time is the time between when a customer places an
order and receives delivery.
• Product variety is the number of different products / configurations
that a customer desires from the distribution network.
• Availability is the probability of having a product in stock when a
customer order arrives.
• Customer experience includes the ease with which the customer
can place and receive their order.
• Order visibility is the ability of the customer to track their order from
placement to delivery.
• Returnability is the ease with which a customer can return
unsatisfactory merchandise and the ability of the network to handle
such returns.
• Supply chain costs affected by network structure:
– Inventories
– Transportation
– Facilities
– Information
Desired Response Time and
Number of Facilities
Figure 4-1 Relationship Between Desired Response Time and Number

of Facilities
Inventory Costs and Number of
Facilities
As the number of facilities in a supply

chain increases, the inventory and
resulting inventory costs also increase
Figure 4-2 Relationship Between Number of Facilities and Inventory

Costs
Transportation Costs and Number
of Facilities
Figure 4-3 Relationship Between Number of Facilities and

Transportation Cost
Facility Costs and Number of Facilities
Facility costs decrease

as the number of
facilities is reduced
Figure 4-4 Relationship Between Number of Facilities and Facility

Costs
Logistics Cost, Response Time, and
Total logistics costs are the

sum of inventory,
transportation, and facility
costs for a supply chain
network.
Figure 4-5 Variation in Logistics Cost and Response Time with Number
of Facilities
A manager must consider the customer needs to be met

and the cost of meeting these needs when designing the
distribution network. Some key customer needs to be
considered include response time, product
variety/availability, convenience, order visibility, and
returnability. Important costs that managers must consider
include inventory, transportation, facilities and handling,
and information. Increasing the number of facilities
decreases the response time and transportation cost but
increases inventory and facility cost.
Design Options for a Distribution
Network (1 of 2)
• Distribution network choices from the manufacturer to the
end consumer
Two key decisions
1. Will product be delivered to the customer location or
picked up from a prearranged site?
2. Will product flow through an intermediary (or
intermediate location)?
Design Options for a Distribution
Network (2 of 2)
One of six designs may be used
1. Manufacturer storage with direct shipping
2. Manufacturer storage with direct shipping and in-
transit merge
3. Distributor storage with carrier delivery
4. Distributor storage with last-mile delivery
5. Manufacturer/distributor storage with customer pickup
6. Retail storage with customer pickup
Manufacturer Storage with
Direct Shipping
In this option, product is shipped directly from the manufacturer to
the end customer, bypassing the retailer (who takes the order and
initiates the delivery request). This option is also referred to as drop-
shipping. The retailer carries no inventory.
Information flows from the customer, via the retailer, to the

manufacturer, and product is shipped directly from the
manufacturer to customers, as shown in Figure 4-6. Online retailers
such as eBags and Nordstrom.com use drop-shipping to deliver
goods to the end consumer. eBags holds few bags in inventory.
Nordstrom carries some products in inventory and uses the drop-
ship model for slow-moving footwear.
Figure 4-6 Manufacturer Storage
with Direct Shipping
Direct Shipping Model
Manufacturer Storage with Direct
Shipping Network (1 of 2)
Table 4-1 Performance Characteristics of Manufacturer Storage with
Direct Shipping Network
Cost Factor Performance

Inventory Lower costs because of aggregation. Benefits of
aggregation are highest for low-demand, high-value
items. Benefits are large if product customization can be
postponed at the manufacturer.
Transportation Higher transportation costs because of increased
distance and disaggregate shipping.
Facilities and Lower facility costs because of aggregation. Some
handling saving on handling costs if manufacturer can manage
small shipments or ship from production line.
Information Significant investment in information infrastructure to
integrate manufacturer and retailer.
Manufacturer Storage with Direct
Shipping Network (2 of 2)
Service Factor Performance
Response time Long response time of one to two weeks because of increased
distance and two stages for order processing. Response time
may vary by product, thus complicating receiving.
Product variety Easy to provide a high level of variety.
Product availability Easy to provide a high level of product availability because of
aggregation at manufacturer.
Customer Good in terms of home delivery but can suffer if order from
experience several manufacturers is sent as partial shipments.
Time to market Fast, with the product available as soon as the first unit is
produced.
Order visibility More difficult but also more important from a customer service
perspective.
Returnability Expensive and difficult to implement.
In-Transit Merge Network
Unlike pure drop-shipping, under which each product in the order is
sent directly from its manufacturer to the end customer, in-transit
merge combines pieces of the order coming from different locations
so the customer gets a single delivery.
Information and product flows for the in-transit merge network are
shown in Figure 4-7. In-transit merge has been used by Dell and can be
used by companies implementing drop-shipping. When a customer
ordered a PC from Dell along with a Sony monitor (during Dell’s direct
selling period), the package carrier picked up the PC from the Dell
factory and the monitor from the Sony factory; it then merged the two
at a hub before making a single delivery to the customer.
Figure 4-7 In-Transit Merge Network
In-Transit Merge Network Cont.
Logistics-service providers (LSP) can consolidate shipments before sending to customer.
Source: https://sloanreview.mit.edu/article/winning-the-last-mile-of-ecommerce/
In-Transit Merge (1 of 2)
Table 4-2 Performance Characteristics of In-Transit Merge

Inventory Similar to drop-shipping.
Transportation Somewhat lower transportation costs than drop-
shipping.
Facilities and Handling costs higher than drop-shipping at carrier;
handling receiving costs lower at customer.
Information Investment is somewhat higher than for drop-shipping.
In-Transit Merge (2 of 2)

Response time Similar to drop-shipping; may be marginally higher.
Product variety Similar to drop-shipping.
Product Similar to drop-shipping.
availability
Customer Better than drop-shipping because only a single
experience delivery is received.
Time to market Similar to drop-shipping.
Order visibility Similar to drop-shipping.
Returnability Similar to drop-shipping.
Distributor Storage with
Carrier Delivery
Under this option, inventory is held not by manufacturers at the
factories, but by distributors/retailers in intermediate warehouses,
and package carriers are used to transport products from the
intermediate location to the final customer.
Amazon and industrial distributors such as W.W. Grainger and

McMaster-Carr have used this approach combined with drop-
shipping from a manufacturer (or distributor). Information and
product flows when using distributor storage with delivery by a
package carrier are shown in Figure 4-8.
Relative to manufacturer storage, distributor storage requires a
higher level of inventory because of a loss of aggregation.
Figure 4-8 Distributor Storage
with Carrier Delivery
Carrier Delivery (1 of 2)
Table 4-3 Performance Characteristics of Distributor Storage with
Carrier Delivery

Inventory Higher than manufacturer storage. Difference is
not large for faster-moving items but can be large
for very slow-moving items.
Transportation Lower than manufacturer storage. Reduction is
highest for faster-moving items.
Facilities and Somewhat higher than manufacturer storage.
handling The difference can be large for very-slow-moving
items.
Information Simpler infrastructure compared to manufacturer
storage.
Carrier Delivery (2 of 2)

Response time Faster than manufacturer storage.
Product variety Lower than manufacturer storage.
Product Higher cost to provide the same level of availability as
availability manufacturer storage.
Customer Better than manufacturer storage with drop-shipping.
experience
Time to market Higher than manufacturer storage.
Order visibility Easier than manufacturer storage.
Returnability Easier than manufacturer storage.
Distributor Storage with Last Mile
Delivery
Last-mile delivery refers to the distributor/retailer delivering the
product to the customer’s location instead of using a package
carrier.
AmazonFresh, Peapod, and Tesco have used last-mile delivery in
the grocery industry. Companies such as Kozmo and Urbanfetch
tried to set up home-delivery networks for a variety of products,
but they failed to survive. The automotive spare parts industry is
one in which distributor storage with last-mile delivery is the
dominant model. It is too expensive for dealers to carry all spare
parts in inventory. Thus, original equipment manufacturers
(OEMs) tend to carry most spare parts at a local distribution
center typically located no more than a couple of hours’ drive
from their dealers and often managed by a third party.
Figure 4-9 Distributor Storage with
Last Mile Delivery
Distributor Storage with Last
Mile Delivery (1 of 2)
Table 4-4 Performance Characteristics of Distributor Storage with Last-
Mile Delivery

Inventory Higher than distributor storage with package carrier
delivery.
Transportation Very high cost given minimal scale economies. Higher
than any other distribution option.
Facilities and Facility costs higher than manufacturer storage or
handling distributor storage with package carrier delivery, but
lower than a chain of retail stores.
Information Similar to distributor storage with package carrier
delivery.
Distributor Storage with Last
Mile Delivery (2 of 2)
Response time Very quick. Same day to next-day delivery.
Product variety Somewhat less than distributor storage with package carrier
delivery but larger than retail stores.
Product More expensive to provide availability than any other option
availability except retail stores.
Customer Very good, particularly for bulky items.

experience
Time to market Slightly longer than distributor storage with package carrier
delivery.
Order visibility Less of an issue and easier to implement than manufacturer
storage or distributor storage with package carrier delivery.
Returnability Easier to implement than other previous options. Harder and
more expensive than a retail network.
Manufacturer or Distributor
Storage with Customer Pickup
In this approach, inventory is stored at the manufacturer or distributor
warehouse, but customers place their orders online or on the phone and then
travel to designated pickup points to collect their merchandise. Orders are
shipped from the storage site to the pickup points as needed.
Examples include 7dream.com and Otoriyose-bin, operated by Seven-Eleven

Japan, which allow customers to pick up online orders at a designated store.
Tesco has implemented such a service in the United
Kingdom, where customers can pick up orders they have placed online.
Amazon is also experimenting with Amazon lockers, where customers can
pick up their shipments. A business-to-business (B2B) example is W.W.
Grainger, whose customers can pick up their orders at one of the
W.W. Grainger retail outlets. Some items are stored at the pickup location,
whereas others may come from a central location.
Amazon Lockers & Tesco
‘Click & Collect’
Figure 4-10 Manufacturer or
Distributor Storage with Customer Pickup
Manufacturer or Distributor Storage
with Customer Pickup (1 of 2)
Table 4-5 Performance Characteristics of Network with Customer
Pickup Sites

Inventory Can match any other option, depending on the location
of inventory.
Transportation Lower than the use of package carriers, especially if
using an existing delivery network.
Facilities and Facility costs can be high if new facilities have to be
handling built. Costs are lower if existing facilities are used. The
increase in handling cost at the pickup site can be
significant.
Information Significant investment in infrastructure required.
Manufacturer or Distributor Storage
with Customer Pickup (2 of 2)
Response time Similar to package carrier delivery with manufacturer or
distributor storage. Same-day pickup is possible for items
stored at regional DC.
Product variety Similar to other manufacturer or distributor storage options.
Product availability Similar to other manufacturer or distributor storage options.
Customer Lower than other options because of the lack of home
experience delivery. Experience is sensitive to capability of pickup
location.
Time to market Similar to manufacturer or distributor storage options.
Order visibility Difficult but essential.
Returnability Somewhat easier, given that pickup location can handle
returns.
Retail Storage with Customer Pickup
In this option, often viewed as the most traditional type of supply
chain, inventory is stored locally at retail stores. Customers walk into
the retail store or place an order online or by phone and pick it up at
the retail store.
Examples of companies that offer multiple options of order placement

include Walmart and Tesco. In either case, customers can walk into the
store or order online. A B2B example is W.W. Grainger: Customers can
order online, by phone, or in person and pick up their order at one of
W.W. Grainger’s retail outlets. Local storage increases inventory costs
because of the lack of aggregation. For fast- to very fast-moving items,
however, there is marginal increase in inventory, even with local
storage.
Retail Storage with Customer Pickup Cont.
Figure 4-11 Retail Storage with
Customer Pickup
Retail Storage with
Customer Pickup (1 of 2)
Table 4-6 Performance Characteristics of Retail Storage with Customer
Pickup Sites

Inventory Higher than all other options.
Transportation Lower than all other options.
Facilities and handling Higher than other options. The increase in handling
cost at the pickup site can be significant for online and
phone orders.
Information Some investment in infrastructure required for online
and phone orders.
Retail Storage with
Customer Pickup (2 of 2)

Response time Same-day (immediate) pickup possible for items
stored locally at pickup site.
Product variety Lower than all other options.
Product More expensive to provide than all other options.
availability
Customer Related to whether shopping is viewed as a positive
experience or negative experience by customer.
Time to market Highest among distribution options.
Order visibility Trivial for in-store orders. Difficult, but essential, for
online and phone orders.
Returnability Easier than other options because retail store can
provide a substitute.
Selecting a Distribution
Network Design
• A network designer needs to consider product
characteristics as well as network requirements when
deciding on the appropriate delivery network.
• The various networks considered earlier have different

strengths and weaknesses.
Selecting a Distribution Network
Design Cont.
Which distribution model is right for which product group?
The working group analyze four product categories for last-
mile distribution:
– Generic products (low risk low value product – MRO, office
supplies)
– Commodities (low risk high value products, basic
production materials such as rubber, sugar, copper)
– Distinctives (high-risk, low-value items and services such
as engineered items, parts)
– Critical items (high risk high value – items critical for final
product such as computer chip)
Selecting a Distribution
Network Design Cont..
Only niche companies end up using a single distribution
network. Most companies are best served by a combination
of delivery networks. The combination used depends on
product characteristics and the strategic position that the
firm is targeting. The suitability of different delivery designs
(from a supply chain perspective) in various situations is
shown in Table 4-8.
Design
Hybrid Network
An excellent example of a hybrid network is that of W.W.

Grainger, which combines all the aforementioned options in
its distribution network.
The network, however, is tailored to match the

characteristics of the product and the needs of the
customer. Fast-moving and emergency items are stocked
locally, and customers can either pick them up or have
them shipped, depending on the urgency.
Design
Hybrid Network - Continued
Slower-moving items are stocked at a national DC and

shipped to the customer within a day or two. Very-slow-
moving items are typically drop-shipped from the
manufacturer and carry a longer lead time. Another hybrid
network is used by Amazon, which stocks fast-moving
items at most of its warehouses and slower-moving items
at fewer warehouses; very-slow-moving items may be
drop-shipped from suppliers.
Delivery Networks for Different Product/
Customer Characteristics (1 of 2)
Table 4-8 Performance of Delivery Networks for Different
Product/Customer Characteristics
Retail Manufacturer Manufacturer Distributor Distributor Manufacturer/

Storage with Storage Storage with Storage with Storage Distributor
Customer with Direct In-Transit Package with Storage with
Pickup Shipping Merge Carrier Last-Mile Customer
Delivery Delivery Pickup
High-demand
product +2 −2 −1 0 +1 −1
Medium-
demand +1 −1 0 +1 0 0
product
Low-demand
Product −1 +1 0 +1 −1 +1
Key: +2 = very suitable; +1 = somewhat suitable; 0 = neutral; −1 = somewhat

unsuitable; −2 = very unsuitable.
Delivery Networks for Different Product/
Customer Characteristics (2 of 2)
Retail Manufacturer Manufacturer Distributor Distributor Manufacturer/
Storage Storage Storage with Storage Storage Distributor
with with Direct In-Transit with with Storage with
Customer Shipping Merge Package Last-Mile Customer
Pickup Carrier Delivery Pickup
Delivery
Very-low-
−2 +2 +1 0 −2 +1
demand product
High product
−1 +2 +1 +1 0 +2
value
Quick desired
+2 -2 −2 −1 +1 -2
response
High product
−1 +2 0 +1 0 +2
variety
Low customer
−2 +1 +2 +2 +2 −1
effort
Key: +2 = very suitable; +1 = somewhat suitable; 0 = neutral; −1 = somewhat
unsuitable; −2 = very unsuitable.
• Distribution networks that ship directly to the customer
are better suited for a large variety of high-value products
that have low and uncertain demand. These networks
incur lower facility costs and carry low levels of inventory
but incur high transportation cost and provide a slow
response time.
• Distribution networks that carry local inventory are
suitable for products with high demand, especially if
transportation is a large fraction of total cost. These
networks incur higher facility and inventory cost but lower
transportation cost and provide a faster response time.
Adding Outbound
Transportation
To the Model
SUP 3033 Supply Chain Network Design

Mehmet BASTI PhD
Supply Chain Network Design: Applying Optimization and Analytics to the Global Supply Chain by
Michael Watson, Sara Lewis, Peter Cacioppi and Jay Jayaraman; published by Pearson Education
Inc., 2013 : ISBN-13:9780133017373
Adding Costs to the Model
■ Transportation is often the most important cost in
a network design study. A retail or manufacturing
company may spend the equivalent of 2% to 3%
of their total revenues on transportation alone.
For a company with $5 billion in revenues, this
could equate to $150 million in spend. You can
imagine that reducing these costs can make a
significant difference to the firm’s performance.
2
■ A change in the number and location of facilities
often impacts transportation costs more than any
other cost in the supply chain. Conversely,
manufacturing, purchasing, or product handling
costs may be fairly constant no matter where
your facilities are.
■ In practice, as with service-level models, many

network design studies consider only
transportation costs.
3
■ We will focus on “outbound” transportation

● This is the cost to ship from the final facility (a
warehouse or plant) to the final customer
● The inbound transportation is the cost to ship
product into the warehouse or plant. We cover
this in later chapter
● We will cover the modes of transportation,
which will apply to both inbound and outbound
4
■ Transportation costs are related to distance, but not completely
● Minimum charges
When you ship a full truckload, there is often a minimum charge
assessed independent of how far you go. That is, a carrier may
charge $1.50 per mile per truckload, with a minimum of $300.
Therefore any shipments within 200 miles will cost the same.
● Regional difference
Transportation rates are not symmetrical due to the imbalance of
supply and demand.
● Different modes of transportation and shipment profiles
In a single network some customers may always order a full truck’s
worth of products at a time, while others may only order a half a
truck. In these cases, the optimization has an incentive to select
facility locations closer to the more expensive half-truck quantity
customers.
5
Formulating Transportation Costs
■ The problem is almost the same
as the practical center of gravity,
except we use the transportation
cost (transi,j) and not distance
■ Objective:
● Minimize total transportation costs
■ Constraints
● Meet all demand
● Open just P warehouses or plants
● If you use a site, you must use it
■ Decisions
● Yi,j– which customer is assigned to
which plant or warehouse. Here we
force this to be 0 or 1, meaning we
single source the customer
● Xi– which plant or warehouse is open
6
Three Important Facts About Transportation
Costs to Help You Understand the Model
■ Demand is expressed in total
■ Transportation costs are expressed in per-unit costs
■ You need a transportation cost for every source-

destination pair you would like to consider
7
Demand is Expressed in Total
■ For each customer point we typically express demand as

the total annual demand
● A customer point is typically a collection of actual ship-
to locations rolled up by geographic proximity (all the
customers in Miami)
● We don’t have to use annual demand, but could use
Quarterly or Half a Year, depending on the
circumstances
8
Demand is Expressed in Total
■ We don’t model every single shipment

● Model would become too large to manage, solve, and
analyze
● Model would not be more accurate anyway– we cannot
forecast next year’s demand at the shipment level
■ But, we can be creative when we set up the customer

points and demand to capture reality
● Break up customers by types of shipments
● Break up customers by importance
9
Transportation Costs are Expressed in
Cost Per Unit
■ If we didn’t express in cost per unit, we would be back to

modeling every shipment
■ Once we have the cost per unit, we can easily scale the
model up or down (with more or less demand)
■ Different trucking and shipping companies will express

their rates differently
● We can convert many different rates to a cost per unit
10
Transportation Costs Per Unit:
Different Modes of Transportation
■ Commercial Truckload of Full Truckload (TL or FTL)

■ Private or Dedicated Fleet
■ Less-Than-Truckload (LTL)
■ Parcel
■ Ocean
■ Rail
■ Intermodal
■ Multi-Stop
Discussion Question:
Why not just use the cheapest mode?
11
Transportation Cost Per Unit:
Converting to Actual Rates to a Cost Per Unit
■ For any transportation structure, you can determine the cost of a load
to go from one location to another and the average shipment size.
This gives us the following formula
■ loadi,j is the cost to ship a load from location i to location j.
■ avgi,j is the average shipment size for that load
12
Transportation Cost Per Unit: Example of
Calculating Average Shipment Size
13
Determining All Possible Transportation
Costs
■ In previous chapters, when we minimized total distance, it

was trivial to build a distance matrix with the distance from
and to every possible point
■ A transportation matrix with all combinations may not be

as simple
● You have cost data on lanes you are using, but not on new ones
● If your carriers give you rates for new lanes, they may be retail
rates and not negotiated rates (so you can’t make fair comparisons
to the rates you have)
● Your data may not be clean enough to use
● You may not have the time to collect this data
Note: A combination of our source and destination is often called a “lane”
14
Costs: Building the Rate Matrix
■ There exist some databases with rates built-in (mostly in

the U.S.)
● UPS or FedEx rates for Parcel
● LTL rates through SMC3 (these tables are often referred to as CZAR)
● Some commercial TL rates, but these may have the problem of being
average rates
■ Use your average per-mile (per-km) rates

■ Use the non-negotiated rates and adjust based on rates
you currently pay
■ Use industry benchmarks
■ Use regression analysis
15
Costs: Sample TL State to State Cost Matrix
16
Costs: Sample Average Cost Calculation
17
Costs: Use Regression to Determine Costs
18
Costs: Practice Regression
■ On the book Web site, in the course info section
in Chapter 6,
(http://networkdesignbook.com/academic-
use/course-materials/) , we have a file for
regression (Raw Truckload Rates for
Regression.xls)
■ This data was from an actual project that built

the rate table using regression
■ Download the file, run a regression, and

interpret the results
Videos about regression analysis
https://www.youtube.com/watch?v=Cltt47Ah3Q4
https://www.youtube.com/watch?v=HgfHefwK7VQ
19
Costs: Multi-Stops
■ In some cases, you deliver to customer making multiple

stops.
● The transportation costs are not point-to-point
■ Although it would be great to pick warehouses and route

at the same time, it is not going to happen
● The size of the WH selection problem is very large (see Chapter 3)
● For every single option in warehouse selections, you have a huge
number of potential routes (see book).
● The combined problem is intractable
● In addition, the scale of the problems is different
− Annual models for network design
− Daily models with time windows for routing
20
Costs: Estimating Multi-Stop Costs
■ However, we still need to estimate the costs for our

network design models
■ High level estimate of the cost

● Use a cost number that is between your TL and LTL costs (otherwise you
likely would not be using multi-stop if did not fall in this range)
● If TL costs are between $0.08 and $0.11 per ton mile and LTL costs are
between $0.31 and $0.35, you can use a cost between these measures
■ Determine an appropriate factor

● See spreadsheet on the book Web site
● This file takes you through the example in the book
■ Run routing optimization iteratively with network design to

refine cost measures
21
Case Study- Baseline
■ Firm ships via UPS

out to customers
■ Currently using one

facility located at
the UPS hub in
Louisville, KY
■ Current cost is
$16.6M
■ Want to consider
other alternatives
22
Case Study- Best 3 Warehouses
■ Cost drop to $15.2M
■ Improved service
● This also improves even
though it is not in the
objective
It is not surprising that the distance to customers

improved. As mentioned previously, transportation costs
are strongly correlated with the distance traveled.
23
Case Study- Best 4 Warehouses
■ Best 4 warehouses,
costs drop to
$15.0M
■ Only $200K in
savings by adding
additional
warehouse
■ This $200,000
difference also tells
us the threshold for
the fixed cost. That is, if the fixed cost of the additional
facility is less than $200,000, it would still
be cost beneficial to open the fourth facility
in Dallas.
Summary- Lessons Learned
■ Transportation costs tend to have the same impact as

minimizing distance– sites are pulled close to the
customers
■ Transportation costs are often the most important aspect

of a network design model
■ Transportation rates come in many different flavors

● You need to carefully build transportation rate structures and make sure
they make sense for your business
● You need to keep in mind the lessons learned from significant digits
■ Adding cost to the model helps you build a better

business case
● Even though the answers may not change
25
Adding Capacity to the
Model
In the previous models, we did not consider capacity. Each

facility could be as large as it needed to be to service the
demand assigned to it. But this may not always be a good
assumption.
Inc., 2013 : ISBN-13:9780133017373
Every Facility in a Supply Chain Has Capacity
■ Why not include it in the model?

● First, demand may naturally be distributed across your facilities.
When you ask the model to open multiple facilities, it will naturally
tend to balance the demand across the facilities to minimize
average distance or cost.
● Second, especially for warehouses, increasing capacity can be
relatively inexpensive. With warehouses, you may be using a third-
party warehouse and extra space simply means asking for it. Even
when you are not using a third party, many warehouses can easily
expand in space.
● Third, often the decisions you are making concern the long term.
And, in that case, today’s capacity is not fixed for the future. You
want the model to return the size of the facility needed. You do not
want to fix this beforehand.
2
Every Facility in a Supply Chain Has Capacity
■ Why include it?
● You are making decisions about expensive plants,

warehouses, or lines
● You are making decisions around highly automated

facilities or equipment
● You are determining number of shifts to run
3
Capacity Examples
■ Swimming Pool Chemicals
■ Warehouse Capacity Utilization

● How do you measure capacity?
■ Paint Company Example
4
Adding Capacity to the Model
■ Deciding on the number to put into the model

● 80% of the true capacity?
● 125% of the measured capacity?
■ For warehouses
● Is the capacity storage?
● Is it inbound put-away, is it outbound shipping? And, what drives
these?
■ For plants
● Is it labor?
● Is it shifts?
● Is it equipment?
5
Difficulty with Capacity Constraints
■ Capacity constraints can cause an infeasible model
● If total available capacity required is more than available
● Can show up if you do not allow enough facilities to open or if the facilities are not
large enough
■ Capacity constraints can cause strange-looking results

● See Figure 5.2. With capacity constraints, Anapolis serves customers that are
closer to another warehouse
■ Capacity constraints can create a very hard problem (see next

slide)
6
Capacity Constraints Can Create a Very Hard
Problem: (It can create a Knapsack Problem)
Single Sourced ■ This single problem is very

Capacity Demand hard, all by itself on realistic-
sized models
10 6 ■ You can see how run time
can increase this if you have
2
to solve this problem for
10 5 every single combination of
3 choices
■ How you can create this
3
problem:
1 ● Tight capacities
● Demand is large relative to
Problem: Assign capacity (if demand is small, there
customers to Total = isn’t as much worry)
warehouses and don’t 20 ● Can happen with warehouses and
single sourcing groups of
split the demand customers
● Can happen with plants and min
lot sizes
Computational Complexity: Another Way
to Look at in addition to the Side Bar
Key Points from the Wikipedia Article
■ “…no known efficient way to locate a solution…”

■ “…no fast solution to them is known…”
■ “…, the time required to solve the problem using any currently known
algorithm increases very quickly as the size of the problem grows.”
● For example, double the size, the run time goes up by 10X
■ “… the time required to solve even moderately sized versions of many

of these problems easily reaches into the billions or trillions of years,
using any amount of computing power available today.”
■ NOTE: Many network design problems happen to run relatively fast.

However, it is not surprising that you can quickly create models that
do not run in a reasonable amount of time
9
Capacity Case Study: Brazilian Distributor
■ Distributor in Brazil with about 100 million units of total demand from
about 25 demand regions
■ Wants to locate 5 warehouses to serve this demand
■ Do not want to split up a demand region
Avg Distance = 326 km
■ The solution above has no capacity constraints

■ They did not like the wide range in sizes of the warehouses and
wanted to put a capacity constraint of 20 million per warehouse
10
Capacity Limit of 20 Million Units
■ If we limit the capacity to 20 million units, the model is no

longer infeasible
■ The demand for Sao Paulo region is 29 million units
■ Since we have a constraint that does not allow us to split

a region, we have a conflict:
● The model cannot both limit the warehouse to 20 million units and serve a
customer with demand for 29 million from a single warehouse
■ In this case, the distributor thought the constraint about

splitting a region was more important
● When you have conflicting constraints, you should carefully consider all
the conflicts to determine the most important ones
11
Capacity Limit of 30 Million Units;
Same Locations
30 million Capacity,
No Capacity
Same Locations
12
Capacity Limit of 30 Million Units;
New Locations
30 million Capacity,
No Capacity
New Locations
13
Summary
■ Capacity constraints impact the model in different way
■ Capacity can be difficult to measure
■ There are many ways to model capacity constraints

● Just one way is presented in the text
■ Capacity constraints can cause your model to

● Give strange results
● Be infeasible
● Run for long time (or just keep running)
14
Alternative Service Levels
and Sensitivity Analysis
Inc., 2013 : ISBN-13:9780133017373
What Does Service Level Mean?
■ Picture yourself standing in the kitchen after a long
week of work or classes. You are starving and
rifling through your delivery menus from a few of
your favorite pizza parlors. Pat’s Pizzeria has
always had your favorite deep dish, but there is
only one location all the way across town and it
always takes at least 60 minutes for your order to
arrive. You then come across the most recent
menu left at your doorstep for a new place called
Primo Pies. You have heard good things about
them and know they have three locations in town,
with one right down the street.
■ You quickly pick up the phone and place your
order for a deep-dish pizza from Primo Pies.
Within 30 minutes you are on the couch enjoying
your pizza and watching your favorite movie. Pure
bliss!
2
■ The rationale you just used to determine which restaurant

to order pizza from is very similar to the decisions
consumers make in almost all product markets.
■ Think back to how Al’s Athletics got their start in the
industry. Locating themselves in a market where the other
big sporting-goods stores couldn’t provide an acceptable
service level led to their success in the industry.
■ For all retailers, this means locating stores close to
customers or locating warehouses that ship online orders
close enough to get next-day service (or be willing to pay
for air freight).
3
■ The reasons customers want stores close is similar to

why businesses want to be able to service these stores
from nearby locations. Al valued being close to customers
in two senses. He wanted his stores in convenient
locations for his customers, but he also wanted his
warehouses close to his stores. Shoppers at Al’s may not
care how far away the warehouses are. But Al knows that
the closer his warehouses are, the faster he can replenish
the stock at stores and ensure that shoppers have the
product that they want when they want it.
4
Supply Chain Network Design & Service Level
■ The term “Service Level” is used too

much without definition
● It has many different meanings
● Unclear definitions do not lead to clear solutions
■ Valid definitions for network design

● Average distance to customers
● Percent of customers within a certain distance
■ Definitions not applicable to network

design
● Fill rate
● Late orders
■ Valid service level constraints

● Maximum distance to a customer
5
Chen’s Cosmetics – Corporate Background
■ Chun Chen, a former makeup artist decided to
team with her brother, a chemist to begin
producing a line of affordable, professional
quality cosmetics to be sold across China. Her
new company was called Chen’s Cosmetics.
■ The company began with a single production
facility in Guangzhou, China which served the
entire customer base made up of distributor’s
warehouse locations across China Existing Distribution from Chen’s
Plant to all Customers
■ The cosmetic line immediately become very popular and Chen’s now finds
themselves with more demand than their existing plant can produce. As the CEO,
Chun Chen decides they will add two new plants in China to alleviate this problem
and offer better service to their customers.
■ After some initial research, Chen knew that building a plant in any of their potential
locations would cost just about the same and decided that their decision for locating
these additional facilities should then be focused on locations which would enable to
the best service to their existing customer base.
■ The question was… how do they determine the location for these additional plants
based solely on Service Level?
Chen’s Cosmetics – China Problem Definition
■ The Logistics team at Chen’s Cosmetics decided to use a commercial Network
Design application to find help them find the optimal answer to their question.
■ The team generated a list of 24 potential locations where there was space and
favorable conditions for building new plants. You can see these locations plotted
as the light blue rectangles with flags on top within the network graphic below.
■ Knowing what they did about the

constant cost of building a plant the
team set out to focus the model on
Service Level, which would require
asking the model to minimize the
weighted average distance to all
customers or maximize the
percentage of customers within a Dark Plant = Existing Guangzhou Location
Lights Plants = Potential Plant Locations
certain distance when selecting the
optimal two plant locations Chen’s Cosmetics Current Network Layered with
24 Potential sites where they could Locate their
additional Plants
Chen’s Cosmetics – China Scenario Results
Objective 1: Offer the best service possible Objective 2: Maximize the demand you
to the customers who purchase the most can service within 800km/2 Day Service
Chen’s Cosmetics – Europe Problem Definition
■ Chen’s Cosmetics does so well in the Chinese market that they find the
opportunity to expand the sale of their products to the European Market as well.
■ The company will continue to produce all products from their three plants in China
but have teamed with successful 3PL in Europe to distribute the imported product
from key warehouse facilities across the country.
■ The 3PL provides them with a list of 48
potential warehouse locations to select
from. The per cubic meter lease cost
would remain the same at all locations,
therefore the team was again looking to
offer the best service level possible with
the three warehouse locations they were
to select.
■ The team also determined that they

would solve the scenario with both Chen’s current European Customer Base layered
with the potential warehouse locations (red
service level objectives, previously triangles) presented to them by the 3PL
discussed and then select the optimal
solution
Chen’s Cosmetics – Europe Scenario Results
Objective 1: Offer the best service possible Objective 2: Maximize the demand you
to the customers who purchase the most can service within 800km/2 Day Service
Selects: Selects:
Paris, France Paris, France
Rome, Italy Belgrade, Serbia
Kremenchug, Ukraine Voronezh, Russia
Service-Level Constraints***
■ Although the decision for Chen’s Cosmetics was simply based on
proximity to customer demand alone, many studies will want to
consider service level as just a part of their entire network
optimization. In this case, service level is no longer the overall
solver objective but a constraint within the larger model.
■ As previously discussed, our first service-level objective asks the
model to locate plants as close to as much demand as possible.
A solution to this model can result in some customer locations
being as close as 20km from the servicing plant while others are
as far away as 4,000km. A modeler on Chen’s team thinks it will
be best for business if the solution ensures that all customers can
be serviced within one-day transit or 400km.
■ After quick analysis of their data, it is clear that this constraint will
cause infeasibility in their model.
Max Distance Example
Figure shows the distance from distributors (customers) to the closest

potential plant. A max distance smaller than these numbers cause
immediate problems.
12
Infeasibility in Network Modeling
■ Infeasibility in modeling can be quite common.
■ When infeasible solutions are determined, the most
common way to pinpoint their cause is to test iterations of
providing the model slightly more freedom in its decisions
until a feasible solution is possible.
■ In this case, Chen’s modeler knows that he has applied
only two major constraints in his model:
1. That only two additional plant locations may be selected for use in
the solution and
2. That all distributors must be located within 1,200 km from one of
the three plant locations.
■ He also knows that the model was feasible prior to adding
the second constraint. Knowing this, he decides to start
by running iterations of the model with increases in the
maximum distance constraint.
13
Max Distance Example
However, 4.10 shows that the interaction can also cause problems.
14
The Importance of Sensitivity Analysis on
Any Solution
■ We have not only learned a good deal about the inclusion
of service levels in a model but also learned the value of
running multiple scenarios to test and further understand
the output and different potential “good” solutions.
Network design models may make assumptions in many

areas:
■ Demand in the future
■ Transportation costs
■ Labor costs
15
The Importance of Sensitivity Analysis on
Any Solution
■ Although intricate forecasting techniques and other

detailed methods of data analysis will help modelers to
create the best assumptions in any model, it is important
that any and all solutions are tested for robustness.
■ This practice is commonly known as sensitivity analysis.

This practice arms the modeler with a better knowledge of
what drives the model and answers. By testing changes in
key variables such as demand, transportation costs, and
labor costs, a modeler is ensuring that an assumption
made in one model input is not dramatically altering the
resultant savings and network recommendations within
the solution.
16
Sensitivity Analysis on Chen’s Cosmetics
Let’s look at an example of sensitivity analysis performed on

Chen’s Cosmetics new distribution structure in Europe. For
demand numbers, Chen’s modeling team uses the estimates from
the marketing team that suggest they will be able to control 65%
of the affordable-cosmetics market in Southwest Europe.
Although Chun Chen has a lot of trust in her marketing

department, she, like all good managers, knows that a forecast is
never accurate. And that forecasting becomes even more difficult
in new markets. So Chun asks the modeling team to run some
sensitivity analysis around the effect this forecast or assumed
market share has on the optimal warehouse location solution.
This way she can rest easy, knowing that the company’s decision
will remain optimal despite the possibility of their underperforming
in comparison to these optimistic market share goals.
17
Sensitivity Analysis on Chen’s Cosmetics
■ The team decides to start their analysis by running the original

model scenario with 15% less demand than was originally
assumed for their South West European region. The solution
recommendation is the same as for that of the original
scenario. Optimal warehouse locations remain in Leipzig,
Marseille, and Odessa. This gives Chen’s the assurance that
their optimal solution is robust enough to withstand up to 15%
less demand in the Eastern European market.
■ The team then continues to test the robustness of their solution

by running scenarios with between 20% and 40% less demand
at 5% intervals. You can quickly see in their solution
comparison report, shown in Figure 4.11, that at 20% less
demand, the solution selects three completely new warehouse
locations in Bucharest, Milan, and Voronezh.
18
Chen’s Cosmetics – Europe Sensitivity Analysis
Chen’s Cosmetics – Europe Sensitivity Analysis
Original Solution
20% Less Demand in Southwest Europe 35% Less Demand in Southwest Europe
Summary
■ There are different measures of service level
● Only some are valid for network design
■ Service levels can be modeled as the objective or as a

constraint
■ When we add these constraints, we can create models

that are no longer feasible
● The solver is faced with contradictory constraints
■ Running many different scenarios is a great way to test

the sensitivity of your solution
21
End-of-Chapter Questions
■ A major consumer products company is making a major
push to improve the number of orders filled on time in full.
This measure is often abbreviated as OTIF. It means that
when a customer places an order for several different
products, you are able to ship all the products requested
in the quantities requested. The company wants you to
run a network design model with this as the sole
objective. Why is a network design study not going to help
this company?
22
■ If you are deciding where to position ambulances to
provide emergency service to residents within a city,
would you rather minimize the average time to each
resident or maximize the percentage of residents within a
certain time? Why?
23
■ If the mathematical formulation only had an objective that
maximized the amount of demand within a certain limit
(say, 800km) and no other objective or restriction on the
distance, how would customers outside of the limit be
assigned to warehouses?
24
Locating Facilities Using Distance
Based Approach:
P-median & P-Center
Inc., 2013 : ISBN-13:9780133017373
1
Facility Location and Allocation
Decisions in Supply Chain Network
Almost every private and public sector organizations has
been faced with the problem of locating facilities at one
time or another in its history. Industrial firms must
determine locations for fabrication and assembly plants as
well as warehouses. Retail outlets must locate stores.
Government agencies must locate offices and other public
services including schools, hospitals, fire stations,
ambulance bases, vehicle inspection stations, and landfills.
In every case, the ability of a supply chain to produce and

market products effectively or deliver high-quality services
is dependent on the location of the facilities in relation to
other facilities and to its supply chain partners.
2
Classification of Facility Location
Problems
Location problems come in a variety of forms, which can be
classified with respect to a number of criteria.
• Continuous space
Universe is defined as a region, such that clients and
facilities may be placed anywhere within the continuum,
and the number of possible locations is uncountably
infinite.
• Discrete space
Universe is defined by a discrete set of predefined
positions.
• Network space
Universe is defined by an undirected weighted graph.
Client positions are given by the vertices. Facilities may be
located anywhere on the graph.
3
Where should the radio tower be located?
• Tower may be positioned anywhere – (continuous)

• Tower may be positioned in five available slots – (discrete)
• Tower may be located on the roadside – (network)
The objective function is the most significant

characterization of a facility location problem.
4
Optimization Methods for
Supply Chain Network Models
■ Optimization is the mathematical discipline which is
concerned with finding the maxima (profit, assembly
line performance, crop yield, bandwidth, etc.) or
minima (cost, loss, risk, etc.) of functions, possibly
subject to constraints.
■ Major field within the discipline of Operations

Research and Management Science
■ Helps
● an organization to get the maximum benefit out of
its resources
● management achieve its goals using scientific
methods.
5
Optimization Methods
■ People use informal “optimization” to make
decisions almost every day.
● Finding the shortest route to school
● Allocating your resources (money, time, etc.)
■ Organizations use formal optimization methods to

address problems across the organization, from
optimal pricing to locating a new facility.
6
Don’t Be Intimidated by the Math
Formulation
■ Going through the math formulation helps you
systematically understand how to approach these
problems
● Just knowing about the math formulation allows you to
think logically about the problem
■ We use the math formulation to help you build

intuition
7
Optimization: Basic Ideas
Optimization Problem Components
● Decision Variables
● Objective Function (to maximize or minimize)
● Constraints (requirements or limitations)
■ Basic Idea
● Find the values of the decision variables that maximize
(minimize) the objective function value, while staying
within the constraints.
8
8
Linear Programming (LP)
■ Objective function and all constraints are linear functions of
the decision variables (e.g., no squared terms, trigonometric
functions, multiplications or ratios of variables).
● LPs are much easier to solve by computer than problems involving
nonlinear functions.
■ Real-world LPs, which contain thousands to millions of

variables, with specialized software.
● Our problems are obviously much smaller, but the basic concepts
are much the same.
Linear functions are those whose graph is a
straight line. A linear function has the following
form. y = f(x) = a + bx. A linear function has one
independent variable and at least one
dependent variable. The independent variable is
x and the dependent variable is y.
9
Solving Network Optimization
Problems
■ Understand the problem; draw a diagram
■ Write a problem formulation in words
■ Write the algebraic formulation of the
problem
● Define the decision variables
● Write the objective function in terms of the
decision variables
● Write the constraints in terms of the decision
variables
10
Definition of “Customers”
■ When modeling a supply chain, we will use the term
“customers” to refer to the end point of the supply chain
● That is, the point we need to deliver to
■ The “customer” need not be an end consumer or even a
single entity.
■ In general, the “customer” can refer to
● Stores
● ZIP or Postal codes you deliver to
● A pool point where goods are consolidated for final shipment to the store
● Another company’s warehouse or manufacturing site
● Your own final set of warehouses
● A port of exit
● And, so on…
■ We have a set of customers we want to serve. We call
this set “J” (the capital letter J). If we use lowercase j, we
are referring to an individual customer.
11
Definition of “Demand”
■ Once we have customers, we need to assign each
customer a “demand”
■ We define the “demand” as the total amount of product
we need to get to each customer in the time horizon of the
model
● Time horizon is usually annual
● Total amount of product is measured in a way that makes sense
for logistics
− Total weight
− Total cases
− Total pallets
− And, so on…
− (But, not total revenue)
■ We refer to demand d (lower case d) and append a
subscript to indicate which customer. So, demand at a
customer is dj.
12
Terminology
■ Facilities: We have a set of facilities we want to pick to
serve customers (these can be warehouses, plants, etc).
We call this set “I” (the capital letter I). If we use
lowercase i, we are referring to an individual facility
■ Distance Matrix: We are concerned about the

distance between facilities and customers. We need a
distance matrix that shows us every combination. We call
this disti,j
● The term dist refers to the fact that this term represents
the distance. The i,j subscript tells us that it is the
distance between facility i and customer j.
13
The Decisions of the Network
Optimization
■ Do we pick a particular facility or not?
● We will designate this decision as Xi.
− Remember the i indicates one of the potential facilities
● We will allow Xi to take on two values, 1 or 0. If it is 1, it means
that this facility is used in the solution. 0 it is not.
● This is a binary variable because it is either 1 or 0. In this case, it
does not make logical sense to open ½ of a warehouse
■ Which customer is assigned to which facility?

● We will designate this decision as Yi,j
− Remember the i indicates one of the potential facilities and j one
of the customers
● In this case, to start, Yi,j can only be 1 or 0. If it is 1, it means that
customer j has all its demand met from facility i.
14
The Objective of the Network
Optimization
■ We want to minimize the total (or average)
weighted distance from the facilities to the
customers in the supply chain.
■ We are using demand as our “weights”
15
The Constraints of the Network
Optimization
■ We have to meet all demand
● If we don’t have this, we could reduce distance by not meeting
demand
■ We can only locate a certain number of facilities. We will

call this number P.
● This allows to us to determine the best 1, 2, 3, and so on facilities
● Without a limit, the model may just select every facility
■ If a customer is served from a facility, then that facility

must be opened
● This constraint ties the X and Y decisions together. This constraint
forces the solver to pick the best values of X and Y at the same
time
16
Key Questions Addressed
By Network Optimization Models
a. How many facilities should be opened?
b. Where should each facility be located?
c. How large should each facility be?
d. How should demand for the facilities’ services
be allocated to the facilities?
The answers to these questions depend intimately on

the context in which the location problem is being
solved and on the objectives underlying the location
problem.
17
Key Questions Addressed
By Network Optimization Models Cont.
The number of facilities to be located as well as the

size of the individual facilities is often a function of
the service/cost trade-offs. In many cases, the quality of
service improves as the number of facilities located
increases, but the cost of providing the service also
increases.
Facility location models are also concerned with the

allocation of demands to facilities. In some cases, it is
important that demands at a site not be split between
facilities.
18
Network Design Problem
(P-median)
● We are given a set of customers with known demands
and a set of potential locations for warehouses.
● We are also given the pairwise distances between
customers and potential facility locations.
● We would like to find the locations of P facilities to
minimize the demand weighted total distance while
serving all customers so that the total transportation
cost is minimized.
● There are no capacity constraints at the facilities.
19
The P-median Problem
■ 1 and 2 facility planar solutions
20
The P-median Problem Cont.
■ p=5, network problem
Tripolis, Greece: 1358
nodes (variable demand
per node; 2256 edges. 5
facility p-median solution.
Colors show demand
allocations (serviced
vertices). Darker lines
show network routes
employed. Note – all
facilities located fairly
centrally
21
The P-median Problem:
Remarks
As there are no capacity constraints at the
facilities, it is optimal to satisfy the demand at a
demand node from a single facility.
An optimal solution can be found by restricting

the search to the demand nodes.
Fixed costs are assumed to be equal.
The number of possible solutions is
22
Size of P-median Problem
■ Suppose you have 25 potential facility locations. If you
want to pick the best location for a single facility, how
many trials you need to make for picking the best
location?
■ How about if you are picking the best locations for 2

facilities? How about 10?
■ What if the potential facility locations is 250?
23
Size of P-median Problem Cont.
• For even moderate values of n and p the number of

possible solutions can be very large:
• For example, if n=20 and p=5, there are 15,504

possible solutions.
• If n=50 and p=10, the problem has more than 1010

possible solutions.
• If you could evaluate 1 million solutions per second, it

will take you 3 hours for total enumeration.
24
Size of P-median Problem Cont..
■ If you are picking multiple sites, the problem size grows

very quickly:
■ You need to use mathematical programming techniques

to sort through these combinations
25
P-median Problem Network
Structure
12 18
C
A
D
15
14
16 E 15
20
18
15
B F
22
12
17
G 20
H
26
Matrix from
Example
…to A
A B C D E
A
B
From
C
B…
D
E
27
Cost Matrix
Site j
Node i
A B C D E F G H
A 0 16 12 29 14 34 28 48
B 16 0 28 33 18 34 12 32
C 12 28 0 18 26 33 40 50
D 29 33 18 0 15 15 30 32
E 14 18 26 15 0 20 15 35
F 34 34 26 15 20 0 22 17
G 28 12 40 30 15 22 0 20
H 48 32 50 32 35 17 20 0
28
P-median Problem Solution
Daskin’s Heuristic Algorithm
29
Selection of First Site
Site j
Node i A B C D E F G H
A 0 16 12 29 14 34 28 48
B 16 0 28 33 18 34 12 32
C 12 28 0 18 26 33 40 50
D 29 33 18 0 15 15 30 32
E 14 18 26 15 0 20 15 35
F 34 34 26 15 20 0 22 17
G 28 12 40 30 15 22 0 20
H 48 32 50 32 35 17 20 0
Total
181 173 200 172 143 175 167 234
Cost
30
Selection of Second Site
Site j
Node i
A B C D E F G H
A 0 16 12 29 14 34 28 48
B 16 0 28 33 18 34 12 32
C 12 28 0 18 26 33 40 50
D 29 33 18 0 15 15 30 32
E 0 0 0 0 0 0 0 0
F 34 34 26 15 20 0 22 17
G 28 12 40 30 15 22 0 20
H 48 32 50 32 35 17 20 0
Total
167 155 174 157 143 155 152 199
Cost
31
Median Sites and Costs
Site No Site Total Cost

1 E 143
2 G 152
Site No Site Assigned Demand Points

1 E A, C, D, E, F
2 G B, G, H
32
Solution Graph
12 C 18
A D
15
14
16 E 15
20
18
15
B F
22
12
17
G 20
H
33
Total Network Cost for
Facilities E and G
Facilities
Demand Points
E G
A 14 0
B 0 12
C 26 0
D 15 0
E 0 0
F 20 0
G 0 0
H 0 20
Column Total 75 32
Total Cost 107
34
Neighborhoods Search for
P-median Problem
35
P-median Problem
Demand
Median Sites and Costs Points
Facilities
E H
12 C 18 A 14 0
A
D B 18 0
15
14 C 26 0
E D 15 0
16 15
18
20 E 0 0
F 0 17
15
B F G 15 0
22
12 H 0 0
17 Column
G 88 17
20 Total
H
Total Cost 105
36
P-median Problem
Median Sites and Demand Facilities
Costs-3 Points A F
A 0 0
B 16 0
12 C 18 C 12 0
A
D
D 0 15
15
14 E 14 0
E F 0 0
16 15
20 G 0 22
18
H 0 17
15
B F
Column
12
22 42 54
Total
17
G
20
H Total Cost 96
37
P-median Sample Problem 2
10 20
C
A
D
15 15
25 20
B F
20
20
10
G
15 E
38
The P-median Problem With
Demand
Demand: 50 Container
Vaughan
20km
Brampton
30km Demand: 100

Container
40km
Toronto
Demand: 120 Container
39
Sample P-Median Problem - 3
100 250
14
A B
13
150
10 E
23 17
C 16
12 Demand
200
D 125
40
Math Formulation
Cannot assign
demands to
an unopened
facility
Number to locate
Each demand
assigned once
Integrality
41
Important Concepts
■ Feasibility
● Feasible Solution: A solution is feasible if all constraints
are satisfied.
● Infeasible Solution: A solution is infeasible if one or
more constraints is violated.
■ Optimal Solution
The optimal solution is the feasible solution with the
smallest objective value
42
Solving Linear Programming
Problems
■ Simplex Method. This is a mathematical approach
developed by George Dantzig. Can solve small problems
by hand.
■ Computer Software. Most optimization software actually
uses the Simplex Method to solve the problems. Excel’s
Solver Add-In and Lindo are examples of such software.
43
Lingo
■ http://www.lindo.com/index.php/ls-downloads/try-lingo
■ Free trial version

■ Solves linear programs
■ Easier to code than Excel
44
Lingo-Download Instructions
Click to the download button that suits your computer's
operating system. When you click, you need to fill in a short
form and then you will receive a link to download the
software. Unzip the installation file, then double click to install
it. When you run Lingo, it will ask you for a license key. Click
on "Demo", then click "OK" in the next window, then click on
"Do Not Create User ID File."
45
Lingo-Modeling
■ To use the algebraic model (Lindo format) directly, you will
do the following: File--> New --> Lindo Model (.ltx) --> OK
■ To open a Lindo file with Lingo, you should open Lingo
first, then File--> Open--> Change Files of type option to
"All Files:--> Click on the Lindo model that you want to
open.
46
Syntax tips for LINGO
■ When you receive a weird error message, check
the rules for syntax in “Help”
■ In the constraints, have only
● numbers on the right hand side
● variables and coefficients on the left hand side
■ Lingo does not accept
● Parentheses
● Constants in the objective function
■ Don’t forget
● ST before constraints
● END after constraints
■ Use
● INT to define Binary integer decision variables
● GIN for General integer decision variables
47
Lingo Code
Minimize 3500YAB + 2000YAC + 2750YAD + 4050YAE + 1400YBA + 4600YBC + 2125YBD + 1950YBE
+ 1000YCA + 5750YCB + 1500YCD + 4200YCE + 2200YDA + 4250YDB + 2400YDC + 2400 YDE +
2700YEA + 3250YEB + 5600YEC + 2000YED
ST
YAA - XA <= 0
YAB - XA <= 0
YAC - XA <= 0
YAD - XA <= 0
YAE - XA <= 0
…
YEA - XE <= 0
YEB - XE <= 0
YEC - XE <= 0
YED - XE <= 0
YEE - XE <= 0
XA +XB + XC+ XD + XE = 1
YAA + YBA + YCA + YDA + YEA = 1

YAB + YBB + YCB + YDB + YEB = 1
YAC + YBC + YCC + YDC + YEC = 1
YAD + YBD + YCD + YDD + YED = 1
YAE + YBE + YCE + YDE + YEE = 1
END
48
Solution of Sample Problem
(P=1)
100 250
14
A B
13
150
10 E
23 17
C 16
12
200 Demand
D
125
49
Solution of Sample Problem (P=1) Cont.
100 250
14 Locate at B
A B for a total
13
150 demand
10 weighted
17 E
23 distance of
10,075
C 16
12
200 Demand
D
125
50
Lingo Results
51
Fixed Cost of Opening A Facility
■ We have not included the fixed cost of opening a facility
■ Instead we used P as a proxy of our budget.
■ Now, let’s assume P is unknown, but we have a fixed cost
associated with each facility. How does the model
change?
■ Let’s try
● Fixed cost = $1000 for each facility
52
Lingo Code for P-Median
Minimize 5000XA + 5000XB + 5000XC+ 5000XD+ 5000XE + 3500YAB + 2000YAC + 2750YAD +
4050YAE + 1400YBA + 4600YBC + 2125YBD + 1950YBE + 1000YCA + 5750YCB + 1500YCD +
4200YCE + 2200YDA + 4250YDB + 2400YDC + 2400 YDE + 2700YEA + 3250YEB + 5600YEC +
2000YED
ST
YAA - XA <= 0
YAB - XA <= 0
YAC - XA <= 0
YAD - XA <= 0
YAE - XA <= 0
…
YEA - XE <= 0
YEB - XE <= 0
YEC - XE <= 0
YED - XE <= 0
YEE - XE <= 0

END
53
Lingo Outputs
54
Capacity of a Facility
■ We implicitly assumed that a facility has infinite
capacity
■ How do we include capacity in the model?
● Assume they all have a capacity of 300
− How many facilities needed?
● Assume they all have a capacity of 500
− How many facilities needed?
55
Lingo Code for P-Median with Capacity
Minimize 5000XA + 5000XB + 5000XC+ 5000XD+ 5000XE + 3500YAB + 2000YAC + 2750YAD + 4050YAE +
1400YBA + 4600YBC + 2125YBD + 1950YBE + 1000YCA + 5750YCB + 1500YCD + 4200YCE + 2200YDA +
4250YDB + 2400YDC + 2400 YDE + 2700YEA + 3250YEB + 5600YEC + 2000YED
ST
YAA - XA <= 0
YAB - XA <= 0
YAC - XA <= 0
YAD - XA <= 0
YAE - XA <= 0
…
YEA - XE <= 0
YEB - XE <= 0
YEC - XE <= 0
YED - XE <= 0
YEE - XE <= 0

100YAA + 250YAB + 200YAC + 125YAD + 150YAE - 300XA <= 0

100YBA + 250YBB + 200YBC + 125YBD + 150YBE - 300XB <= 0
100YCA + 250YCB + 200YCC + 125YCD + 150YCE - 300XC <= 0
100YDA + 250YDB + 200YDC + 125YDD + 150YDE - 300XD <= 0
100YEA + 250YEB + 200YEC + 125YED + 150YEE - 300XE <= 0
END 56
Lingo Output
57
Supply Chain Network Design
Set Covering Models***
■ In some location problems, a maximum distance
exists a priori. (covering distance)
■ For example:
● Some businesses guarantee service within a pre-
determined time (e.g. , 20 minute pizza delivery).
■ Demand within the covering distance of its closest
facility is considered covered.
■ Being closer to a facility than the maximum
distance does not improve satisfaction.
58
Set Covering Models Cont.
P-center
problems
1-center
Minimize the maximum distance to the facility
59
Set Covering Models Cont..
P-center
problems
p-center
Minimize maximum distance to the closest facility
60
The P-centre Problem
● Broadly, locate p facilities to service demand at n>p
locations, such that the maximum distance travelled by
customers is minimised
● This is a form of ‘coverage’ problem – attempting to
provide facilities for all customers in a manner which
ensures that (i) every customer can been serviced and (ii)
travel distances/times/costs are not excessive
● Max travel time is often computed from (network)
distance. If a maximum acceptable value is specified then
may require a lot of facilities (i.e. p becomes a variable)
61
The P-center Problem
■ Given P facilities, find the minimum coverage
distance such that all demands are covered
■ In effect, we are minimizing the maximum
demand from any customer to the nearest facility
■ No matter how insignificant, the most “remote”
demand node drives the solution
■ Can be solved as a linear problem – but how to
incorporate a max function?
62
The P-Center Problem Cont.
■ P-centre: p=5, network problem
Tripolis, Greece: 1358

nodes (variable
demand per node;
2256 edges. 5 facility
p-centre solution.
Colours show demand
allocations (serviced
vertices). Darker lines
show network routes
employed. Note -
much wider spread of
facilities vs the p-
median solution
63
The Comparison of
P-Center Problem & P-Median Problem
P-Center Problem P-Median Problem

The P-Center Problem
Graphical Example
Locate at A: E
14
A B is __km away
13
10 E Locate at B: C
23 17
is __km away
C 16
12
Locate at D: A
D is __km away
65
Graphical Example Cont.
110 300
14
A B
13
85
10 E
23 17
C 16
12 Fixed Cost
225 D 90
66
Graphical Example Solution
Node A covers nodes A, B, C, and D with coverage dist = 22

14
A B
13
10 E
23 17
C 16
12
D
67
Graphical Solution
Node A covers nodes ?? with coverage dist = 15

14
A B
13
10 E
23 17
C 16
12
D
68
Graphical Solutions
Node A covers nodes A, B and C with coverage dist = 15

14
A B
13
10 E
23 17
C 16
12
D
69
The P-Center Problem Graph
If maximum coverage distance 15 km how many facilities we

need? Which locations?
14
A B
13
10 E
23 17
C 16
12
D
70
Generic Set Covering
Solutions
10
A D
8 7
12
9 C 13 F
13
9
7
B E
17
P-center Solutions
10
A D
8 7
12
9 C 13 F
13
9
7
B E
17
P-center and Set Covering
Set covering P-center

Sample P-Median Problem - 4
50
70
10
Sarnia Barrie
10
80
15 Ajax
15 20
Windsor 15
10 Demand
50 Brampton
110
74
Introduction to Location
Analysis Methods
Chapter 2
Inc., 2013 : ISBN-13:9780133017373
1
Factors to Consider
in Supply Chain Network Design
• Understanding a specific business’s context
• Interdependencies with other related supply chains
• Master agreements within an organization
• Country-specific logistics infrastructure
• Factors that influence location decisions .
Facility Location
■ Three most important factors in real estate:

1. Location
2. Location
3. Location
■ Facility location is the process of identifying the

best geographic location for a service,
distribution or production facility.
Location Strategy
■ Infrequent decision based on:

● Demand outgrowing existing capacity.
● Local changes in labor productivity, exchange rates,
costs, local attitudes.
● Shifts in demographics and customer demands.
■ Location options:
● Don’t move, expand an existing facility.
● Maintain current sites, add another facility.
● Close an existing facility and move to another location.
Factors That Affect Location
Decisions
■ General factors.
■ Global Region or Country decision.
■ Sub-region or state decision.
■ Community/site decision.
General Factors
■ Globalization.
■ Market (customer) proximity
● High population areas, close to JIT partners
■ Suppliers proximity
● Transportation costs, perishability, bulk
■ Labor proximity and productivity
● Proximity—local wage rates, unions, special skills
availability
● Productivity—low cost may be linked to low productivity
and vice versa)
■ Competitor proximity
● Clustering—due to a major resource in the area).
Global Region or Country
■ Key International Locations
● North America, Europe, Pacific Rim.
■ Key Considerations
● Political/legal concerns.
● Cultural issues (including business).
● Infrastructure: supplies, communication, utilities.
● International trade issues.
− Exchange rates.
● Market access issues.
● Labor availability, attitudes, productivity, costs.
● Quality-of-life issues.
Sub-Region or State
■ Key factors:
● Government incentives.
● Corporate desires
● Market and demographic factors.
● Proximity to raw materials and customers
● Attractiveness of region (culture, taxes, climate, etc.)
● Environmental regulations of state and town.
● Economic conditions.
● Costs of key inputs and advertising media.
● Cost and availability of utilities.
● Labor availability, costs, attitudes towards unions.
Community / Site
■ Factors:
● Financial incentives.
● Site size and cost.
● Transportation options and costs.
● Utility options and costs.
● Nearness of services/supplies needed.
● Legal climate and community receptiveness.
− Zoning restrictions
● Environmental concerns.
■ Significant trends:
− Moving to the suburbs
− Industrial parks
− Moving closer to end user (JIT)
Location Analysis Techniques
Analysis should follow 3 step process:
Step 1: Identify dominant location factors
Step 2: Develop location alternatives
Step 3: Evaluate locations alternatives
• Location factor rating

• Center-of-gravity
• Load-distance
10
Location Factor Rating Method
A popular approximate method used to compare the
attractiveness of several alternatives against several
quantitative and qualitative dimensions.
The main steps to follow include:
• Identifying the factors;
• Assigning weights to each factor (0,00–1,00; all weights add to 1);
• Subjectively rating each alternative against each factor
(0–100; 100 being the best);
• Multiplying the factor rating by the weight, then adding the weighted
scores; and
• Selecting the alternative with the highest total weighted score.
Location Factor Rating (example)
SCORES (0 TO 100)
LOCATION FACTOR WEIGHT Site 1 Site 2 Site 3
Labor pool and climate .30 80 65 90
Proximity to suppliers .20 100 91 75
Wage rates .15 60 95 72
Community environment .15 75 80 80
Proximity to customers .10 65 90 95
Shipping modes .05 85 92 65
Air service .05 50 65 90
Weighted Score for “Labor pool and climate” for

Site 1 = (0.30)(80) = 24
12
Location Factor Rating
WEIGHTED SCORES
Site 1 Site 2 Site 3
24.00 19.50 27.00
20.00 18.20 15.00
9.00 14.25 10.80 Site 3 has the
11.25 12.00 12.00 highest factor
6.50 9.00 9.50 rating
4.25 4.60 3.25
2.50 3.25 4.50
77.50 80.80 82.05
13
Location Factor Rating With Excel
Supplement 7-14
The Center of Gravity Approach is Common in
Everyday Life and Is a Simple Starting Point for a
Supply Chain Problem
Center of gravity for network design is determining the location

that minimizes the weighted average distance to all the demand
points.
■ The simplest facility-location problem is the center of gravity (COG)
problem.
■ COG problems are good learning tools
■ The examples we saw in Chapter 1 can be quite complex. The COG
models allow us to build intuition before moving on to those problems
15
The Centre-of-Gravity Method
■ The Center of Gravity Method is an approach that seeks
to compute geographic coordinates for a potential single
new facility that will minimize costs. It’s an approach
where the main inputs that it considers are the following:
● Markets
● Volume of goods shipped
● Shipping costs
■ This method is beneficial because it’s

● Simple to compute,
● Considers existing facilities,
● Minimizes costs.
16
Centre-of-Gravity Method
In general, transportation costs are a function of
distance, weight, and time. The center-of-
gravity, or weight center, technique is a
quantitative method for locating a facility such
as a warehouse at the center of movement in a
geographic area based on weight and distance.
This method identifies a set of coordinates
designating a central location on a map relative
to all other locations.
17
Centre-of-Gravity Method Cont.
The starting point for this method is a grid map set up on
a Cartesian plane. This Involves mapping a number of
locations on an x, y coordinate grid and then finding a
central location closest to all.
x-coordinate = sum of (each location’s x-coordinate * weight)

total weight
y-coordinate = sum of (each location’s y-coordinate * weight)

total weight
https://www.wikihow.com/Calculate-Center-of-Gravity
Grid-Map Coordinates (Problem 1)
y n n
∑ xiWi ∑ yiWi
2 (x2, y2), W2 i=1 i=1
y2 x= n y= n
∑ Wi ∑ Wi
1 (x1, y1), W1 i=1 i=1
y1
where,
x, y = coordinates of new
3 (x3, y3), W3 facility at center of gravity
y3 xi, yi = coordinates of existing
facility i
Wi = annual weight shipped
from facility i
x1 x2 x3 x
19
Centre-of-Gravity Method (Problem 1)
y
A B C D
700
C x 200 100 250 500
600 (135) y 200 500 600 300
B
W 75 105 135 60
500 (105)
Miles
400
D
300
A (60)
200 (75)
100
0 100 200 300 400 500 600 700 x

Miles x-coordinate = sum of (each location’s x-coordinate * weight)
total weight
y-coordinate = sum of (each location’s y-coordinate * weight)

total weight
20
Centre-of-Gravity Method (Problem 1:
Solution)
n
∑ xW
i i
i=1 (200)(75) + (100)(105) + (250)(135) + (500)(60)
x= = = 238
n 75 + 105 + 135 + 60
∑ W
i
i=1
n
∑ yW
i i
i=1 (200)(75) + (500)(105) + (600)(135) + (300)(60)
y= = = 444
n 75 + 105 + 135 + 60
∑ W
i
i=1
21
Centre-of-Gravity Method (Problem
1: Solution Cont.)
y
A B C D
700
C x 200 100 250 500
600 (135) y 200 500 600 300
B
W 75 105 135 60
500 (105)
Center of gravity (238, 444)
Miles
400
D
300
A (60)
200 (75)
100
0 100 200 300 400 500 600 700 x

Miles
Center-of-Gravity With Excel
Formula for
x coordinate
Supplement 7-23
Physics Weighted-Average
Centering (Problem 2)
■ Consider a hypothetical country based on the
principles of rational organization. Let’s call this
country Logistica. The citizens of Logistica must
choose a location for their capital.
■ Although all the citizens wish to live as near the
seat of government as possible
24
Let’s Build Our Intuition For How
Optimization Solves These Problems
Where Should the Citizens of Logistica Locate Their Capital?
Map of Logistica and Population by City
25
Option #1: Right in the Center
The planners of Logistica first decide to place the capital in the

most central location possible. Initially, they choose their
capital location by considering the boundaries of Logistica, and
then selecting the location that centers the country
geographically.
Option #1: Right in the Center Cont.
• The population of Logistica is not uniformly distributed

across its interior. In general, there are more people in the
East than the West, with a relatively empty middle. Thus, a
geographically centered capital would incur pointlessly large
travel times for everyone.
• The map shows a square representing the approximate
location of the geographic center. The planners calculate
that this point is, on average, 471 miles from each citizen.
27
Option #1: Right in the Center Cont.
Distance Calculation
We can estimate the distance in miles between any two
latitude and longitude points that are approximately
halfway between the equator and the North or South Pole
with the following equation:
Dist(miles)ab is the distance from point a to point b. Longa,

Lata, Longb, and Latb are the longitudes and latitudes
expressed as decimal numbers of point a and point b,
respectively. If you want this measure to be in kilometers
instead of miles, you simply change the 69 to 111.
28
Option #1: Right in the Center
Distance Calculation in Excel
EXCEL Method
http://www.spatialthoughts.com/blog/gis/calculate-distance-spreadsheet/
29
Option #2: Pick The Weighted
Average Longitude and Latitude
The mathematical formula for physics center of gravity (or
the weighted average location) would find the coordinates
of Logistica’s capital as follows.
30
Option #2: Pick The Weighted
Average Longitude and Latitude Cont.
Here, Lon represents a city’s longitude, Lat

represents its latitude, and P represents a city’s
population. Here we are using the population as
the weighting factor. In other problems, many
different weighting factors can be used. In
network problems, customer demand is the most
common. These formulas compute the weighted
average of the longitude and latitude of all the
existing cities (represented by the set C).
31
Option #2: The Physics Center of
Gravity
The physics center of gravity for Logistica using the
previous Lon and Lat equations is shown in Figure.
What else can go wrong with this approach?

32
Option #2: The Physics Center of
Gravity Cont.
With this approach, the weighted-average distance a
citizen must travel is down to 388 miles.
However, the location happens to fall
in shark-infested waters offshore from
a mountainous and deserted region.
Obviously, the citizens of Logistica are
not pleased with the resulting
proposal. Though it’s good for a
physics calculation, the planners
quickly realize that the center of gravity
calculation would not prove practical.
Logistica has its share of uninhabited coastlines,
mountainous regions, treacherous canyons, and forest
preserves. These would not be good locations for the
capital. The planners realize that they need to exclude
these locations as options.
Reasons for Why The Physics Center of
Gravity Calculations Should Not Be Used
• It does not minimize what you want to minimize: weighted
average distance.
• It might point to the middle of a large lake, an ocean, the
top of a tall mountain, a wildlife preserve.
• It will almost never land in a location that exploits existing
infrastructure, such as a city, a population center with a
workforce, railways, highways, or ports.
• The method cannot take advantage of true road
distances or travel restrictions—it must rely on straight-
line estimates based on latitude and longitude.
• It cannot be extended to include factors like costs,
capacities, different types of facilities, different products,
multiple levels of facilities (hub and spoke, suppliers,
warehouses, retailers, etc.), or other practical
considerations.
Practical Center of Gravity
Some planners start to ask the question of
what happens when they simply pick the
three Eastern cities of 5, 16, and 11
35
A Better Solution: Minimize Average
Distance
Picking Cities 5, 16, and 11
36
We Can Use Each of the Cities As a
Potential Location for the Capital
Note that there is nothing to prevent us from picking other

locations for the capital. The methodology would be the
same, we would just need the latitude and longitude of the
new location
The Source of this picture is the Logistica Excel file
37
Lessons Learned from Center of
Gravity Problems
■ In this chapter, we learned that locating a single point
relative to demand is best done by minimizing the
average weighted distance. You can accomplish this by
picking your solution from a candidate list of locations.
■ When picking one point, it is often simple enough to list
out all the combinations and pick the best one. In the end,
when you have a solution, it will tend to pull the single
point close to as much demand as possible.
38
Lessons Learned from Center of
Gravity Problems Cont.
■ Although the physics concept of a weighted-average
central point is appealing to our intuition and
mathematically easy, it is flawed for network design
studies. It can lead to undesirable locations as it is really
minimizing the population multiplied by miles squared.
■ Also, it cannot be extended to take advantage of existing

infrastructure, road distance, or other facility costs.
39
Load-Distance Technique
A variation of the center-of-gravity method for determining
the coordinates of a facility location is the load-distance
technique. In this method, a single set of location
coordinates is not identified. Instead, various locations are
evaluated using a load-distance value that is a measure
of weight and distance.
Steps of Load-Distance Technique;
■ Compute (Load x Distance) for each site
■ Choose site with lowest (Load x Distance)
■ Distance can be actual or straight-line
40
G
100 TL 360 TL
60 TL
DC
Brampton
41
Load-Distance Calculations
n
LD = ∑ ld i i
i=1
where,
LD = load-distance value
li = load expressed as a weight, number of trips or units
being shipped from proposed site and location i
di = distance between proposed site and location i
di = Sqrt( (xi - x)2 + (yi - y)2 )
where,
(x,y) = coordinates of proposed site
(xi , yi) = coordinates of existing facility
42
Load-Distance
Potential Warehouse Customers
Site X Y A B C D
1 360 180 X 200 100 250 500
2 420 450 Y 200 500 600 300
3 250 400 Wt 75 105 135 60
Compute distance from each site to each customer

Site 1 dA =Sqrt((xA - x1)2 + (yA - y1)2)= Sqrt((200-360)2 + (200-180)2 )= 161.2
dB = Sqrt((xB-x1)2 + (yB-y1)2 ) = Sqrt((100-360)2 + (500-180)2 ) = 412.3
dC = 434.2 dD = 184.4
43
Load-Distance Cont.
Site 2 dA = 333 dB = 323.9 dC = 226.7 dD = 170
Site 3 dA = 206.2 dB = 180.3 dC = 200 dD = 269.3
Compute load-distance
n
LD = ∑ ld i i
i=1
Site 1 = (75)(161.2) + (105)(412.3) + (135)(434.2) + (60)(184.4) = 125,063
Site 2 = (75)(333) + (105)(323.9) + (135)(226.7) + (60)(170) = 99,789
Site 3 = (75)(206.2) + (105)(180.3) + (135)(200) + (60)(269.3) = 77,555*
* Choose site 3
44
Load-Distance With Excel
=B7*C11+C7*C12+D7*C13+E7*C14
Supplement 7-45
Load-Distance Example -2
Potential Warehouse Customers
Site X Y A B C D
1 300 150 X 150 120 350 400
2 320 250 Y 100 400 450 300
3 150 300 Wt 100 120 90 80
Compute distance from each site to each customer
46
Load-Distance Example -2
47
A Load-Distance Model Example 3
■ Starwood Inc. is considering where to locate its
warehouse in order to service its four Ohio stores located
in Cleveland, Cincinnati, Columbus, Dayton.
■ Two sites are being
considered;
W=15
Mansfield and
Springfield, Ohio.
Use the load-
distance model to W=4
make the decision. W=10
W=12
■ Given the analysis presented in this chapter, where would
you put the capital of Logistica and why? What factors
went into your decision?
■ Besides the weighted-average distance and the
percentage of customers within a certain distance of the
capital, what other factors might the citizens of Logistica
want to consider? Of these factors, which are quantifiable,
and which are qualitative?
■ Name a reason why minimizing weighted-average
distance is more important than maximizing the
percentage of customers within a certain distance. Now,
name a reason why maximizing the percentage of
customers within a certain distance is more important
than why minimizing weighted-average distance.
49
Center of Gravity Problem
■ Given the following assumptions below, what are the
coordinates for the new potential DC location?
■ Let’s assume the following:
■ Customer 1 has a daily demand goods volume of 2,500 units
And the current coordinates of the existing facilities:
Customer 3
Customer 2
Customer 1
50
Strategic Network Design
Source: Francas and Simon, 2011, Camelot Management Consultants
1
Three reasons to have strategic
network design on the agenda:
• Good network design means higher profitability: effective
network design optimizes investments in capital-intensive
assets, lowers operational costs, and helps to reduce working
capital in the supply chain while maintaining the targeted
customer service.
• Good network design means management of growth:
comprehensive network design determines where and when
new facilities are needed to ensure that new markets can be
accessed effectively, and new products can be manufactured
as needed.
• Good network design means less business risks: our business
world is continuously changing. An optimal supply chain
footprint reduces the supply chain’s exposure to supply and
demand risks by creating agility and flexibility at strategic level.
Strategic network design ensures a
competitive supply chain
• The strategic value of network design
• Network design is the strategic planning process

for evaluating alternative structures for a supply
chain and selecting the one that maximizes
profitability and helps to improve performance at
each link in the supply chain.
• Network design plays a crucial role for companies

in all industries striving to deliver outstanding
supply chain performance.
Network design must ensure the
efficiency of global and often
complex operations
Key characteristics of today´s
competitive markets
A holistic approach
to network design,
which is tightly
aligned to the
overall supply chain
and corporate
strategy, is required
to ensure success
in today’s
demanding
markets.
If delivered successfully, strategic network design
improves a company’s supply chain performance
in a variety of areas:
• Higher return on assets as network design optimizes capital-
intensive investments that are typically of a highly irreversible
nature.
• Lower supply chain cost as network design sets the
conditions for ensuring operational efficiency of virtually all
supply chain-related functions and processes.
• Reducing the supply chain’s exposure to supply and
demand risks.
• Creating strategic preconditions for responding to changes
in the business environment and enabling future growth.
• Higher customer service and lower time-to-market for new
products through lower lead times throughout the supply
chain.
Strategic decisions made at network level affect all
levels of supply chain management and provide the
framework for successful tactical and operational supply
chain processes
Opening a new production plant triggers
various decisions at all decision levels
Different drivers for strategic
network design
There are many drivers
that can prompt
companies to initialize
network design projects.
Nevertheless,
considering the
constantly changing
economic environment
and the significant
contribution of an
optimally designed
supply chain to overall
business success,
companies are well
advised to
continuously evaluate
the performance of their
networks and place
supply chain review and
design on their strategic
agenda.
Typical business questions answered
through strategic network design
As shown in
Figure, a well-
structured network
design approach
provides optimal
answers regarding
supplier base,
manufacturing
footprint and
distribution
systems, and
establishes the
basis for far-
reaching network
design decisions.
Aligning network design with
strategy
• The business strategy must be the ultimate
starting point for network design
• Network design translates an envisioned

business strategy into the supply chain.
Therefore, you must ensure that your decisions
regarding network design are driven by your
company’s strategic direction.
Key strategy dimensions
that affect network design
• As shown in Figure, there are four
key strategy dimensions which
primarily characterize such
considerations.
• Strategic network design makes a
significant contribution to achieving
the desired performance in each
dimension. Network design affects
the cost dimension through
optimizing both capital and
operational expenditures, and the
service dimension, for example
through providing appropriate
customer service levels. Internal
throughput times as well as lead
times to customers affect
performance with regard to time,
while agility captures the capability
of a supply chain to respond to
changing requirements.
Objectives and reference models
for strategic network design
It is crucial to derive the

concrete objectives for
network design based on
the envisioned business
strategy. While you might
consider different objectives,
you should focus on the
objectives that support the
intended business strategy.
Objectives for network
design can be either
financial targets, business
targets, or social &
environmental targets
Reference models for strategic
network design
When defining the strategic direction and objectives for a supply
chain, a key question for companies is: which principles should
guide their strategic network design? Industry benchmarks as well
as state-of-the-art research studies point out that companies are
well advised to follow two main design models: Lean network
design and agile network design
The following factors characterize
an agile network design:
• Structural flexibility, e.g. through flexible plants
and flexible workforce models, allows you to respond
to demand fluctuations or supply chain disruptions. A
flexible manufacturing footprint also forms an
efficient hedge against currency and cost risks.
• Surge capacity at the plant and supplier level
ensures that the supply chain can provide high
customer service when demand increases.
• Operational flexibility, such as multi-sourcing
strategies and flexible transshipment modes, makes
supply chain processes capable of dealing with
short-term changes in operational requirements.
Lean network design focuses
on the following aspects:
• Lead-time reduction is characterized by
optimizing internal throughput times and lead
times to customers.
• Simplification reduces network complexity, for
instance by concentrating products or
processing steps at certain facilities and
outsourcing the processing steps that add less
value.
• Leveled capacity allocation eliminates
capacity bottlenecks and helps to avoid over-
investment in assets and low utilization rates.
Key decisions for strategic
network design
In order to ensure that the whole supply chain
follows strategy and objectives and to identify the
right scope for a network design project, it is
essential to take an integrated and comprehensive
view on the whole supply chain, including key
suppliers, the various manufacturing facilities, all
warehouses, transportation channels, and the end
customers. This “cradle to grave” approach in terms
of geographical and functional scope can help
companies to achieve and maintain outstanding
performance in their supply chain in the long run.
network design-2
network design-3
As shown in Figure, network design can help to
define the optimal number and location of facilities
(manufacturing plants, warehouses, and distribution
centers), the allocation of capacity and technology
requirements to facilities, the assignment of products
to plants, and the flow of goods throughout the
supply chain. Furthermore, network design should
determine whether services and processes should
be conducted internally or externally. For example,
network design projects can be used to decide
whether products should be manufactured in-house
or by contract manufacturers, considering all of the
implications for the supply chain.
Need to balance different
objectives
To reach optimal decisions, you typically need to
balance different objectives. For instance, the
trade-offs between conflicting cost objectives
affect the optimal degree of centralization in a
company´s distribution network. Companies
aiming to decide on the optimal number of
distribution centers have to ensure that all
relevant cost dimensions are adequately
incorporated, in order to minimize total supply
chain costs while maintaining or improving
delivery service levels.
Key success factors for
strategic network design
The ultimate goal of a

strategic network design
project is to gain
competitive advantages
and secure long-term
success by providing the
right structure for the
supply chain
The network design cycle
The strategic network design
process
Generally, the strategic network design process

should run through a structured sequence of
planning steps
The main phases of the
network design process
• Definition of objectives and network strategy:
Identifying all relevant targets in alignment with the
overall supply chain strategy, ranking them according to
their relevance, and investigating appropriate trade-offs
are extremely important prerequisites for a successful
network design project.
• Analysis of the supply chain: An in-depth and future-
focused analysis of the as-is supply chain using
information and data from IT systems as well as from
key stakeholders helps to identify current strengths and
weaknesses and to refine objectives.
network design process-2
• Scenario generation: Alternative supply chain
scenarios are developed from a business perspective.
To elaborate the business scenarios, optimization tools
can be used in this phase, for example to decide on
optimal facility locations or capacity levels.
network design process-3
• Scenario evaluation: The developed scenarios
undergo a validation and prioritization process. Both
quantitative and qualitative factors must be considered
when ranking the identified scenarios according to their
benefits and business feasibility.
• Selection of supply chain design: A detailed
business case and implementation plan is essential for
the final selection of a certain scenario. Typically, this is
only done for a few scenarios which have been
evaluated as the best ones.
The Value of Supply Chain
Network Design
Chapter 1
Inc., 2013 : ISBN-13:9780133017373
1
Course Learning Outcome 1.1
Describe the purposes of network design for

an organization
Learning Outcomes
• Understand how supply chain design facilitates network
integration.
• Explain why and when network redesign is required.
• Know what the different stages in network design and
implementation entail.
• Apply the main network design and implementation steps.
• Identify the major factors to be taken into account during
network design and in locating facilities.
• Know which modelling approaches and techniques are
applicable for network design.
Supply Chain Introduction
Plants, Warehouses, Customers
factories, ports
suppliers
■ The physical supply chain is the collection of facilities

such as plants, factories, suppliers, warehouses, and
ports needed to get product from the source to the final
point of consumption
4
What is Network Design?
Plants, factories, Warehouses, Customers
suppliers ports
■ The discipline used to determine the optimal location and

size of facilities and the flow through the facilities is called
supply chain network design
● How many plants should we have?
● How many warehouses should we have?
● How should product flow?
■ Network design is sometimes called network modeling
● You build a mathematical model of your supply chain
● You solve the model using optimization techniques
● You then pick the best answer
5
Why Network Design?
•Supply chain network design can have a major impact an
organisation’s competitiveness.
• Network provides the veins, or channels, through which
inventory flows.
• Main network activities are movement, storage and
transformation. Extends from the supply sources to the
consumers of final products.
• Decisions relating to the supply chain network have a
significant effect for a relatively long period of time.
•80 percent of a supply chain’s cost structure is determined
by network design.
What is Supply Chain Network?
7
A Visual Look at Network Design
In this case, the green dots are our customers. The rectangles in
Mexico, Iowa, and Delaware are the plants. The red triangles in Iowa
and Delaware are existing warehouses. The black triangles represent
potential warehouses. This network design problem is about picking
the right number and locations of warehouses to best serve the
customers.
8
Network Integration & Supply Chain Design
•Network integration requires the right design and
configuration.
•Involves the structural dimensions of the network
as well as the careful selection of the members in the supply
chain.
• Relates to decisions regarding customer service; inventory
policy; transportation modes; and the location and size of
stocking points.
•Key considerations include availability of sources
of supply, manufacturing facilities and where the target
markets are located.
•Consider and evaluate a number of alternative supply
chain networks (cost benefit trade-offs).
Stage-gate Model for Supply Chain Design
Value of Supply Chain Network Design:
Mergers and Acquisitions– Mars and Wrigley
■ In April of 2008, Mars and Wrigley announced their merger
● From the press release (emphasis added): “The true value of this
combination arises primarily from enhanced prospects for growth.
The merger will generate a new world of opportunities for our
people, in addition to the potential for cross-pollination of ideas and
brands and further enhancements of sales, marketing and
distribution infrastructures.”
■ In the Fall of 2009, Mars and Wrigley spoke at the national CSCMP
event (www.cscmp.org) about their supply chain network design project
and the merging of their distribution infrastructures
● $10 million in initial savings from just modeling the baselines of the
two networks (in other words, uncovering some quick savings)
● Management in a better position to make decisions for the new
supply chain
● Project complete in approximately 3 months
11
Merger and Acquisitions– Whirlpool and Maytag
■ In March 2006, Whirlpool completed its acquisition of
Maytag
■ In Sept 2008, DC Velocity magazine reported on

combining the two supply chains
● Total savings target of $400 million

● $40 million of the savings would come from reduced
freight and warehousing (through optimized network
design)
● Ran the project internally
12
Merger and Acquisitions– Miller and Coors
■ Miller and Coors merged operations in the US to become
MillerCoors in the summer of 2008
■ They promised investors $750 million in savings
■ They frequently reported on the supply chain network

modeling results
● “Supply chain integration continues to proceed on schedule. The brewery
optimization project is nearing completion, as product moves are more
than 90 percent complete.”--- May 2010
● “In the third quarter, MillerCoors successfully completed initial product
transitions within its national brewery network. The company will continue
to focus on further network optimization through peak/non-peak season
sourcing changes, as well as opportunities for increased efficiencies.” ---
Nov 2010
13
Supply Chain Transformation at The Home Depot
■ Went from supplier to stores to
using warehouses (RDCs)
■ Helped create a more
competitive supply chain
■ Projected a $1 billion decrease
in inventory
14
Contingency Planning at Chemical Company
■ Global chemical company

■ Political unrest in Egypt in
2011 caused them to close
their plant with 24-hour
notice
■ Used network design to
quickly figure out where to
make product formerly
made in Egypt in order to
continue to meeting all
demand
■ Advantage of using a tool in
this situation is minimizing
disruption
■ Corona Virus?
15
Active Learning
■ Find some examples how Covid-19 changed supply

chain networks?
■ 5 Minutes
16
Classic Case and $5M in Savings
Before After
The red triangles are warehouses

The brown rectangles are plants
The green dots are customers
The blue lines show warehouse to customer assignments
■ Currently they had plants and distribution on each coast,

shipping just their product
■ When they optimized their territories, they had to ship
product across the country to get it to the right
warehouse. But the overall savings was around $5M
with shorter trips to the customers
17
Production Sourcing for Pepsi
■ Pepsi Bottling Group (PBG), now part of PepsiCo,
wanted to optimize which plants should make which
product
● They wanted to use network design technology to
consider the entire supply chain when making this
decision
■ They were able to achieve significant results:

● An increased number of cases available to sell due to
reduced warehouse out of stocks
● Reduction in raw material and supplies inventory from
$201 million to $195 million
● A 2 percentage point decline in the growth of transport
miles even as PBG revenue grew
● Increase in the return on invested capital
18
Modeling a Closed-Loop Supply Chain
Source: IBM Case Study
■ Johnson Controls runs a closed-loop battery supply chain

● The plants ship new batteries to the retailers
● The retailers sell the new and collect the used batteries
● The used batteries are recycled
● The recycled material is sent to the plants and is an important input for new
batteries
■ Johnson Controls still must answer network design
questions:
● Which suppliers and facilities should be used to optimize transportation costs?
● How much capacity should be contracted with each supplier and facility?
● Which battery plant and distribution center should serve which customers?
19
As We’ve Seen, You Can Answer a Wide
Range of Questions with Network Design
■ How many warehouses should we have, where should
they be, how large should they be, what products will they
distribute and how will we serve our different types of
customers?
■ How many plants or manufacturing sites should we have,
where should they be, how large should they be, how
many production lines should we have and what products
should they make, and which warehouses should they
service?
■ Which products should we make internally and which
should we source from outside firms?
■ If we source from outside firms, which suppliers should
we use?
20
You Can Answer a Wide Range of
Questions with Network Design
■ What is the trade-off between the number of facilities and
overall costs?
■ What is the trade-off between the number of facilities and the
service level?
■ How much does it cost to improve the service level?
■ What is the impact of changes in demand, labor cost, and
commodity pricing on the network?
■ When should we make product to best manage and plan for
seasonality in the business?
■ How do we ensure the proper capacity and flexibility within the
network? To meet demand growth, do we need to expand our
existing plants or build new plants? When do we need to add
this capacity?
■ How can we reduce the overall supply chain costs?
Why Supply Chain Network Design
Strategy?
■ Search for ways to lower cost

■ Improve customer service
■ Competitive advantage
■ Usually conflicting goals

● Understand where the main emphasis is
22
The Need for Long-Range Planning
■ In the short run, a firm’s supply chain network

and the locations of its key facilities are fixed.
■ Changing facility locations in the short run can

be impractical due to: Site availability, leases,
contracts, and investments
■ In the long run, however, the design of the

overall network must be thought of as variable.
23
To Answer These Questions and To See
the Value from Our Cases, We Need Data
■ Three broad classes of network design data
● Geographic data
● Warehouse data
● Plant data
■ We will then solve the network design problem using

mathematical optimization
● This is a proven method for this type of problem and
allows us to sort through all the variations
■ Finally, we will need to discuss the non-quantifiable

aspects of network design
24
Active Learning!
■ What is optimization?
■ What is mathematical optimization?
■ What kind of optimization we make in our daily life?
■ 5 Minutes!
25
Geographic Data Introduction
26
Why Does Geography Matter In These
Problems?
■ Transportation Costs
■ Service Levels
■ Input costs: labor, utilities, raw materials
■ Skills available
■ Risk
■ Taxes
27
How Location Affects The System
Freight Freight
Rates Volumes
Facility
Location
Freight Service
Costs Levels
Product Sales
Costs Volumes
Why Even Have Warehouses?
Why Have Plants (or multiple plants)?
Warehouses Plants
• Consolidate product • Service levels
• Buffer lead time • Transportation costs
• Service Levels • Economies of Scale
• Store inventory pre-build • Taxes
• Transportation mode • Split up steps in the
trade-off production process
Types of Warehouses: Types of Plants:

Distribution centers, mixing center, Manufacturing, Assembly, supplier, co-
cross-docks, central and regional, hub packer, toller, contract manufacturer
and spoke, 3rd part or owned, etc.
29
Network Design and Optimizations
Determine the basic structure of the supply chain
Goal
Identify the network design option
creating lowest total-cost for supply
chain!
30
Response Time and
Number of Warehouses
■ For example, Borders provides its customers with books on the

same day but requires about 400 locations achieve this goal for
most of the United States. Amazon, in contrast, takes about a week
to deliver a book to its customers, but uses only about six locations
to store its books.
31
Transportation Costs and
● Inbound transportation costs are the costs incurred in
bringing material into a facility. Outbound transportation
costs are the costs of sending material out of a facility.
● Outbound transportation costs per unit tend to be higher
than inbound costs because inbound lot sizes are typically
larger. For example, the Amazon warehouse receives full
truckload shipments of books on the inbound side, but
ships out small packages with only a few books per
customer on the outbound side.
● Increasing the number of warehouse warehouses
decreases the average outbound distance to customers,
thus decreases the outbound transportation costs.
32
● As long as inbound transportation economies of
scale are maintained (Inbound lot sizes are
typically larger), increasing the number of facilities
decreases the total transportation cost.
Transportation ● If the number of facilities is increased to a point

Costs where inbound lot sizes are also very small and
result in a significant loss of economies of scale in
inbound transportation, increasing the number of
facilities increases total transportation cost.
Number of warehouses
33
Transport cost as a function of
distribution warehouses
Total
Transport
Transport
Inbound
Cost
Outbound
34
Facility Costs and
Facility costs decrease as the number
of facilities is reduced because a
Facility consolidation of facilities allows a firm
Costs to exploit economies of scale.
35
Inventory and Number of Warehouses
To decrease inventory costs, firms try to

consolidate and limit the number of facilities
Inventory
in their supply chain network. For example,
with fewer facilities, Amazon is able to turn
its inventory about 12 times a year, whereas
Borders, with about 400 facilities, achieves
only about two turns per year.
36
Inventory
Network and Number
Inventory of Warehouses
Cost Minimization
Average Inventory as a Function of Number of

Warehouse warehouses
Total Costs Related to
Total Costs
Total Costs
Facilities
Inventory
Transportation
38
Service Level and Number of Warehouses
Customer service level in a supply chain is a

function of several different performance indices.
The first one is the order fill rate, which means
customer demands that are met from stock.
Response Stockout rate is the complement of fill rate and
Time represents orders lost due to a stockout. Another
measure is the backorder level, which is the number
of orders waiting to be filled. To maximize customer
service level, one needs to maximize order fill rate,
minimize stockout rate, and minimize backorder
levels. Another measure is the probability of on-
time delivery, means customer orders that are
fulfilled on-time, i.e. within the agreed-upon due date.
Number of
Warehouses
39
Logistics Costs and Response Time with
Response Time Each firm should have at least the

number of warehouses that
minimize the total logistics costs
Total Logistics Costs
40
When would opening more warehouses
make sense?
Response Time Operating beyond that point makes

sense if the revenue generated from
better responsiveness exceeds the cost
of better responsiveness.
Total Logistics Costs
41
Network Design Problems are
Solved with Optimization
■ Major Components of Optimization

● Objectives– what is it you want the math program to optimize
(either minimize or maximize)? The objective function allows us to
compare one solution to another and determine which one is
better
● Constraints– what are rules of a legitimate solution? What are

the limits on capacity, what can be made where, and so on.
Constraints guide us to find solutions that are feasible
● Decision Variables– what decisions is the model allowed to

make? In network design, this is typically the facilities picked, the
assignments, and what is made where
● Data– what data goes into the model?
42
Optimization Example
■ We have $5 and we would like to buy 1 Coffee, 1 Muffin

and 1 Cookie it should come in 10 minutes.
■ Objective: Buy these three items with lowest cost but
satisfying constraints
■ Constraints: Budget, Time (10 minutes), Items
■ Decision: Which Brand you buy this is your decision.
■ Data: Google Maps(for distance), Store Web Sites for
Price and availability of products.
4min
5min
3min 15min
$4.7
$3.9
$12 $4.5
43
A Few Notes About Data
■ Precision vs Accuracy
■ Don’t Forget Significant Digits
■ Remember how well you can predict the future (demand, oil, currency)
44
A Few Notes About Data -2
■ In the cartoon, the extra precision on the left actually

makes things worse for our poor analyst (who is about to
be hit by a piano). The analyst has to spend too much
time trying to understand the data and misses the
opportunity to take the much-needed action of getting out
of the way. In network modeling, time horizon for making
decisions is much longer, but the data is also much more
complex. Extremely detailed analysis of data causes the
project team to miss their opportunity to impact the supply
chain in a positive way.
■ Our goal for collecting data for a network design model is

to define the data needed with the right level of
significance to make the relevant decisions.
45
Qualitative Data You Must Consider
■ Strategy
■ Risk
■ Disruption Costs
■ Willingness to Change
■ Public Relations/Image
■ Competitors
■ Supply chain Partners
46
Supply Chain Network Design
•Supply chain network may be relatively simple or very

complex.
• Supply chain attributes that determine supply chain
complexity: customer base; product portfolio; supplier base;
manufacturing (process, scale and variety); logistics scope
(inbound and outbound); and systems/applications used.
•Supply chain network design is generally not a regular
activity (carried out every few years).
• Capacity utilization, throughput and inventory turnover:
indicators used to evaluate how effective supply chain design
is.
Supply Chain Network Design (cont.)
Key determinants of supply chain network structure:

Markets
• Marketing strategy and competitive requirements
• Market segments, customers targeted, anticipated demand
Facilities
• Number, size and location
• Geographic placement of production facilities, stocking points and transfer
facilities
Sources of supply
• Location and number of sources of material, goods and services
• Standardisation of major equipment, spare parts
Transport
• Physical link between customers and suppliers
• Enabling the flow of materials and resources
• Transport decision’s major impact
Who is Involved in a Network Design
Project?
■ Operations
● Can include transportation, warehousing,
manufacturing, and vendors
● Provides information on how the business runs and
what is possible
■ Sales/Marketing
● Provides information on service level constraints and
evaluates how different scenarios will impact sales
● Provides forecast data
■ Finance
● Determines cost inputs and validates cost of scenarios
■ Information Technology
● Assists with data collection
49
Why Strategic Planning is Difficult
■ Crosses boundaries within an organization

● Sales, Finance, Operations– multiple plants may be
multiple profit centers
■ Data
● A lot of data can be required
− Customer Demand, Location and Service Requirements
− Transportation (inbound, outbound, FTL, LTL, Air, Ship,
duties/tariffs)
− Warehousing and production information
● New types of data
− Data for new alternatives
− Data viewed differently than standard accounting
■ Tough Optimization problems
50
Why Strategic Planning is Difficult -2
■ Decisions made with remaining uncertainty

● Policies that cannot be quantified
● Labor and Supplier Constraints
● Future demand, future prices
● Political stability, currency fluctuations
● Ability to shape the data in a strategic study: Costs,
demand, constraints
■ Implementation
● Changes suggested can be disruptive
● Changes can be radical
51
Design for Robustness
•Supply chains encounter some unforeseen disruptions,
unplanned events or things that might go wrong (e.g. strikes,
breakdowns and storms).
•Risk assessment and contingency planning; methods to
deal with such eventualities.
•Ensure that the network does not perform poorly under
other-than-expected conditions.
• Some deviations from the original assumptions:
• What if marketing was wrong about demand?
• What if the cost of supply doubles?
• What seemed like a good idea given forecasted conditions may kill
the company under other conditions.
End-of-Chapter Questions -1
■ ABC Bottling Company’s sales have been expanding

rapidly. Their single plant, which ships directly to
customers, is now out of capacity. What factors should
they consider when they decide whether to expand the
existing plant or build another one? If they build another
plant, what factors should they consider when they locate
this plant?
- Response time?
- Transportation Cost?
- Service Level?
- Inventory cost?
- Competition?
- Sales?
- Facility Cost?
53
End-of-Chapter Questions -2
■ A small medical supply company in Australia has just

developed a never before seen product with major pre-
release orders from around the globe already. This
company will need more production capacity to support
their forecasted sales for this new blockbuster product. If
they simply expand their plant in Australia, they estimate
that their production, transportation, and warehousing
costs will be approximately $450 million (AUD).
54
End-of-Chapter Questions -2 Cont.
■ After a careful network design study, they have found two

solutions that people in the company generally like.
■ Solution #1: Estimated cost of $375 million with a new large
plant in China to supplement their existing plant in Australia.
■ Solution #2: Estimated cost of $385 million with three new
smaller plants in China, Brazil, and Italy to supplement their
plant in Australia. These plants would service their local
regions.
■ (Assume the costs listed here include all the costs that are
relevant.) What would be the best reasons for picking
solution #1? For picking solution #2? Why is it important for
this firm to consider other nonquantifiable factors when
determining their best course of action for expansion?

Greta All Merged File

Uploaded by

Copyright:

Available Formats

You might also like

Greta All Merged File

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Greta All Merged File

Uploaded by

Copyright:

Available Formats

Supply Chain Management: Strategy,

Planning, and Operation

1. Focusing exclusively on unit cost rather than total cost

Performance Dimension Activity Affecting Performance Impact of Offshoring

Supply chain visibility Scheduling and expediting Poorer visibility

Raw material costs Sourcing of raw material Could go either way

Product returns Quality Increased returns likely

Inventories Lead times, inventory in transit Increase

Table 6-3 Supply Chain Risks to Be Considered During Network

Category Risk Drivers

Delays High capacity utilization at supply source Inflexibility

Forecast risk Inaccurate forecasts due to long lead times,

Table 6-3 [Continued]

Category Risk Drivers

• Good network design can play a significant role in

Table 6-4 Risk Mitigation Strategies During Network Design

Risk Mitigation Strategy Tailored Strategies

Increase capacity Focus on low-cost, decentralized capacity for

Get redundant suppliers More redundant supply for high-volume products,

Increase responsiveness Favor cost over responsiveness for commodity

Table 6-4 [Continued]

Risk Mitigation Strategy Tailored Strategies

Limited flexibility configuration, where every plant produces at least

• The term value chain means that at each stage of the

– Outbound logistics refers to all activities from the point of a

– Procurement is obtaining purchased inputs, such as

5. Develop adequate flexibility in corporate decision

Source: Kotabe and Helsen

The establishment of what to do and

The achievement and retention of global

Source: Michigan State University

• Taxes on Corporate Income

• Offset Trade and Local Content

Source: Global Supply Chain Associates (GSCA)

• Spare Parts Logistics:

Return on Assets (ROA) is an indicator of how profitable a company is

• Redesign the entire worldwide supply chain

Source: Global Supply Chain Associates (GSCA)

Source: Global Supply Chain Associates (GSCA)

• Distribution – the steps taken to move and store a

• Companies in the same industry often select very

These PC companies have chosen three

Figure 4-1 Relationship Between Desired Response Time and Number

As the number of facilities in a supply

Figure 4-2 Relationship Between Number of Facilities and Inventory

Figure 4-3 Relationship Between Number of Facilities and

Facility costs decrease

Figure 4-4 Relationship Between Number of Facilities and Facility

Total logistics costs are the

A manager must consider the customer needs to be met

Information flows from the customer, via the retailer, to the

Cost Factor Performance

Cost Factor Performance

Service Factor Performance

Amazon and industrial distributors such as W.W. Grainger and

Cost Factor Performance

Service Factor Performance

Cost Factor Performance

Customer Very good, particularly for bulky items.