T2 Facility Decisions

IEDA 3460 Tutorial 2
Designing the Supply Chain Network

Facility Decisions
Junxue ZHANG
jzhangfv@connect.ust.hk, Office 5569
September 19/26, 2022
IEDA 3460 Tutorial 1 September 19/26, 2022 1 / 32

T2: Facility Decisions
▶ Design supply chain structure:

facilities (network), inventory, transportation
▶ A framework for network design decisions:
• Phase I: Network structure design
• Phase II - IV: Location & sizing of facilities
Case study 1: LP formulation

Case study 2: a budget problem
▶ Design under uncertainty:
Case study 3: managing growth at SportStuff.com

Phase I: Network structure design
▶ Facilities: production sites (e.g. factories) and storage sites (e.g

warehouses).
▶ Design purpose: achieve a strategic fit, balancing the service level
(for meeting customer needs) and the cost (of meeting needs).
Customer needs: response time, product availability, ...
Cost: inventory, transportation, facility and handling, ...
▶ Decide for the trade-off between the responsiveness and efficiency.

An example of trade-off: number of facilities v.s. logistic cost
Figure 1: Inventory cost Figure 2: Facility cost

Figure 3: Transportation cost

Draw a graph for total logistic cost vs. number of facilities.

(Tip: total logistic cost = inventory cost + transportation cost + facility
cost)

Other examples of trade-offs:

▶ delivery / pickup?
▶ flow through an intermediary (or intermediate location)?

Example of distribution networks
Figure 4: Manufacturer Storage with Direct Shipping

Figure 5: In-Transit Merge Network

Figure 6: Distributor Storage with Carrier Delivery

Figure 7: Distributor Storage with Last-Mile Delivery

Figure 8: Manufacturer or Distributor Warehouse Storage with Consumer

Pickup

Phase II-IV: Location & sizing of facilities
▶ Phase I: Network structure design

▶ Phase II: Decide on regional facility configuration
▶ Phase III: Decide on suitable sites (identify geographic location for a
site within a region)
▶ Phase IV: Decide on plant location and capacity allocation
Location & sizing of facilities: a balance between cost and demand
satisfaction.

Demand Allocation Model
▶ Decision variable: xij - quantity from plant i to market j

▶ Objective function: minimize variable costs (c)
▶ Constraints: meeting the demands without exceeding the capacity
n X
X m
min cij xij
i=1 j=1
Xn
s.t. xij = Dj
i
m
X
xij ≤ Ki , xij ≥ 0
j

Capacitated Plant Location Model
▶ Decision variable:
xij - quantity from plant i to market j
yi - whether plant i is open
▶ Objective function: minimize fixed costs (f ) + variable costs (c)
▶ Constraints: meeting the demands without exceeding the capacity
n
X n X
X m
min fi yi + cij xij
i=1 i=1 j=1
Xn
s.t. xij = Dj , xij ≥ 0
i
m
X
xij ≤ Ki yi , yi ∈ {0, 1}
j

Multiple sourcing:
n
X n X
X m
Min fi yi + cij xij
i=1 i=1 j=1
Xn
s.t. xij = Dj
i
m
X
xij ≤ Ki yi
j
Figure 9: Multiple sourcing yi ∈ {0, 1}

Single sourcing:
n
X n X
X m
Min f i yi + cij xij Dj
i=1 i=1 j=1
Xn
s.t. xij = 1
i
m
X
xij Dj ≤ Ki yi
j
Figure 10: Single sourcing yi , xij ∈ {0, 1}

Gravity Methods for Location
▶ Decision variable: location (x, y )

▶ Objective function: minimize the total cost from (x, y ) to
(xi , yi ), i = 1, 2, ..., n
Fi - cost of shipping one unit for one mile to location i
Di - quantity to be shipped to location i
p
▶ Constraints: distance di = (x − xi )2 + (y − yi )2
n
X
min d i Di F i
i=1

Case study 1: LP formulation
Blue Computers, a major server manufacturer in the U.S. currently has

plants in Kentucky and Pennsylvania. The Kentucky plant has a capacity
of 10 million units a year and the Pennsylvania plant has a capacity of
15 million units a year. The firm divides the U.S. into 5 markets:
Northeast, Southeast, Midwest, South and West. The firm wants to
build a new plant with a capacity of 15 million units per year and
abandon one of the present plant. Potential sites being considered are
in North Carolina and California.
Q: If Blue Computers has the objective of minimizing total fixed and
shipping costs, where should it build its plant and what should its
shipping plan be?

Linear programming in Python
How to solve LP/NLP (i.e. Integer Programming)?

▶ Algorithms for LP (Simplex methods, ...)
▶ “Solver” function in Excel
▶ “PulP” package in Python

Case study 2: a budget problem
Lily plans to budget for snacks (brownies, ice cream, cola and cheese
cake). Her problem is
min c1 x1 + c2 x2 + c3 x3 + c4 x4
s.t. 400x1 + 200x2 + 150x3 + 500x4 ≥ 500 (at least 500 calories)
3x1 + 2x2 + 0x3 + 0x4 ≥ 6 (at least 6g chocolate)
2x1 + 2x2 + 4x3 + 4x4 ≥ 10 (at least 10g sugar)
2x1 + 4x2 + 1x3 + 5x4 ≥ 8 (at least 8g fat)
x1 , x2 , x3 , x4 ≥ 0
Q: What are optimal quantities, x1∗ , x2∗ , x3∗ , x4∗ ?

Review III: Design in an uncertain environment
Uncertainty in decision environment:

▶ demand
▶ prices
▶ exchange rates
▶ competition
▶ ...

Review III: Design in an uncertain environment
Being flexible in an uncertain environment can be beneficial (for meeting

the uncertain demand) or costly.
▶ new product flexibility: PC industry, ...
▶ volume flexibility: automotive industry, ...
▶ mix flexibility: consumer electronics industry, high-end fashion, ...
Appropriate flexibility is an effective approach for a global supply chain
to deal with a variety of risks and uncertainties.
Example: flexible plants vs. dedicated plants (Lec 2, Page 44)

Case Study: SportStuff.com
Managing Growth at SportStuff.com

In December 2008, Sanjay Gupta and his management team were busy
evaluating the performance at SportStuff.com over the previous year.
Demand had grown by 80 percent. This growth, however, was a
mixed blessing. The venture capitalists supporting the company were
very pleased with the growth in sales and the resulting increase in
revenue. Sanjay and his team, however, could clearly see that costs
would grow faster than revenues if demand continued to grow and
the supply chain network was not redesigned. They decided to
analyze the performance of the current network to see how it could be
redesigned to best cope with the rapid growth anticipated over the
next three years.

Sanjay Gupta founded SportStuff.com in 2004 with a mission of supplying

parents with more affordable sports equipment for their children. Parents
complained about having to discard expensive skates, skis, jackets, and
shoes because children outgrew them rapidly. Sanjay’s initial plan was for
the company to purchase used equipment and jackets from families and
any surplus equipment from manufacturers and retailers and sell these
over the Internet. The idea was well received in the marketplace, demand
grew rapidly, and by the end of 2004, the company had sales of $0.8
million. By this time, a variety of new and used products were being sold,
and the company received significant venture capital support.

In June 2004, Sanjay leased part of a warehouse in the outskirts of St.

Louis to manage the large amount of product being sold. Suppliers sent
their product to the warehouse. Customer orders were packed and
shipped by UPS from there. As demand grew, SportStuff.com leased
more space within the warehouse. By 2007, SportStuff.com leased
the entire warehouse and orders were being shipped to customers
all over the United States. Management divided the United States into
six customer zones for planning purposes. Demand from each customer
zone in 2007 was as shown in Table 1.

Zone Demand in 2007 Zone Demand in 2007

Northwest 320,000 Lower Midwest 220,000
Southwest 200,000 Northeast 350,000
Upper Midwest 160,000 Southeast 175,000
Table 1: Regional Demand at SportStuff.com for 2007
Sanjay estimated that the next three years would see a growth rate of
about 80 percent per year, after which demand would level off.

Sanjay and his management team could see that they needed more
warehouse space to cope with the anticipated growth.
▶ Option 1: lease more warehouse space in St. Louis.
▶ Option 2: lease warehouses all over the country.
Warehouses leased could be either small (about 100,000 sq. ft.) or large
(200,000 sq. ft.). Small warehouses could handle a flow of up to 2
million units per year, whereas large warehouses could handle a flow of
up to 4 million units per year. The current warehouse in St. Louis was
small. The fixed and variable costs of small and large warehouses in
different locations are shown in Table 2.

Location Fixed Cost ($/year) Variable Cost ($/unit flow)

Seattle 300,000/500,000 0.20/0.20
Denver 250,000/420,000 0.20/0.20
St. Louis 220,000/375,000 0.20/0.20
Atlanta 220,000/375,000 0.20/0.20
Philadelphia 240,000/400,000 0.20/0.20
Table 2: Fixed and Variable Costs of Potential Warehouses (Small/Large)
The inventory holding cost

√ per year for a single warehouse is by
observation about $600 F , where F is the number of units flowing
through the warehouse per year. To approximate, the inventory holding
cost is $475, 000Y + 0.165F . For each facility, Y = 1 if the facility is
used, 0 otherwise.

SportStuff.com charged a flat fee of $3 per shipment sent to a

customer. An average customer order contained four units.
SportStuff.com in turn contracted with UPS to handle all its outbound
shipments. UPS charges were based on both the origin and the
destination of the shipment and are shown in Table 3.
NW SW UM LM NE SE
Seattle $2.00 $2.50 $3.50 $4.00 $5.00 $5.50
Denver $2.50 $2.50 $2.50 $3.00 $4.00 $4.50
St. Louis $3.50 $3.50 $2.50 $2.50 $3.00 $3.50
Atlanta $4.00 $4.00 $3.00 $2.50 $3.00 $2.50
Philadelphia $4.50 $5.00 $3.00 $3.50 $2.50 $4.00
Table 3: UPS Charges per Shipment (Four Units)

Questions:
▶ What is the cost SportStuff.com incurs if all warehouses leased are
in St. Louis?
▶ What supply chain network configuration do you recommend for
SportStuff.com? Why?
▶ How would your recommendation change if transportation costs
were twice those shown in Table 3?

Reference
”Supply Chain Management: Strategy, Planning, and Operation”

Chopra & Meindl (C & M), 6th Ed.
▶ Facility decision: Chapter 3, Supply Chain Drivers and Metrics
▶ Network design: Chapter 4-6

T2 Facility Decisions

Uploaded by

Document Information

Original Description:

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

T2 Facility Decisions

Uploaded by

Copyright:

Available Formats

IEDA 3460 Tutorial 2

Designing the Supply Chain Network

September 19/26, 2022

IEDA 3460 Tutorial 1 September 19/26, 2022 1 / 32

▶ Design supply chain structure:

Case study 1: LP formulation

IEDA 3460 Tutorial 1 September 19/26, 2022 2 / 32

▶ Facilities: production sites (e.g. factories) and storage sites (e.g

IEDA 3460 Tutorial 1 September 19/26, 2022 3 / 32

An example of trade-off: number of facilities v.s. logistic cost

Figure 1: Inventory cost Figure 2: Facility cost

IEDA 3460 Tutorial 1 September 19/26, 2022 4 / 32

Figure 3: Transportation cost

IEDA 3460 Tutorial 1 September 19/26, 2022 5 / 32

Draw a graph for total logistic cost vs. number of facilities.

IEDA 3460 Tutorial 1 September 19/26, 2022 6 / 32

Other examples of trade-offs:

IEDA 3460 Tutorial 1 September 19/26, 2022 7 / 32

Figure 4: Manufacturer Storage with Direct Shipping

IEDA 3460 Tutorial 1 September 19/26, 2022 8 / 32

Figure 5: In-Transit Merge Network

IEDA 3460 Tutorial 1 September 19/26, 2022 9 / 32

Figure 6: Distributor Storage with Carrier Delivery

IEDA 3460 Tutorial 1 September 19/26, 2022 10 / 32

Figure 7: Distributor Storage with Last-Mile Delivery

IEDA 3460 Tutorial 1 September 19/26, 2022 11 / 32

Figure 8: Manufacturer or Distributor Warehouse Storage with Consumer

IEDA 3460 Tutorial 1 September 19/26, 2022 12 / 32

▶ Phase I: Network structure design

IEDA 3460 Tutorial 1 September 19/26, 2022 13 / 32

▶ Decision variable: xij - quantity from plant i to market j

IEDA 3460 Tutorial 1 September 19/26, 2022 14 / 32

IEDA 3460 Tutorial 1 September 19/26, 2022 15 / 32

Figure 9: Multiple sourcing yi ∈ {0, 1}

IEDA 3460 Tutorial 1 September 19/26, 2022 16 / 32

Figure 10: Single sourcing yi , xij ∈ {0, 1}

IEDA 3460 Tutorial 1 September 19/26, 2022 17 / 32

▶ Decision variable: location (x, y )

IEDA 3460 Tutorial 1 September 19/26, 2022 18 / 32

Blue Computers, a major server manufacturer in the U.S. currently has

IEDA 3460 Tutorial 1 September 19/26, 2022 19 / 32

How to solve LP/NLP (i.e. Integer Programming)?

IEDA 3460 Tutorial 1 September 19/26, 2022 20 / 32

Q: What are optimal quantities, x1∗ , x2∗ , x3∗ , x4∗ ?

IEDA 3460 Tutorial 1 September 19/26, 2022 21 / 32

Uncertainty in decision environment:

IEDA 3460 Tutorial 1 September 19/26, 2022 22 / 32

Being flexible in an uncertain environment can be beneficial (for meeting

IEDA 3460 Tutorial 1 September 19/26, 2022 23 / 32

Managing Growth at SportStuff.com

IEDA 3460 Tutorial 1 September 19/26, 2022 24 / 32

Sanjay Gupta founded SportStuff.com in 2004 with a mission of supplying

IEDA 3460 Tutorial 1 September 19/26, 2022 25 / 32

In June 2004, Sanjay leased part of a warehouse in the outskirts of St.

IEDA 3460 Tutorial 1 September 19/26, 2022 26 / 32

Zone Demand in 2007 Zone Demand in 2007

IEDA 3460 Tutorial 1 September 19/26, 2022 27 / 32

IEDA 3460 Tutorial 1 September 19/26, 2022 28 / 32

Location Fixed Cost ($/year) Variable Cost ($/unit flow)

The inventory holding cost

IEDA 3460 Tutorial 1 September 19/26, 2022 29 / 32

SportStuff.com charged a flat fee of $3 per shipment sent to a

IEDA 3460 Tutorial 1 September 19/26, 2022 30 / 32