Download as pdf or txt
Download as pdf or txt
You are on page 1of 19

Civil Engineering Decision and Systems Analysis

CE 31600

Introduction to Civil Engineering Systems Analysis


Excerpts from Chapter 1 and Chapter 2 of Civil and Environmental Systems Engineering (Revelle et
al.) and Introduction to Operations Research (Hillier and Lieberman)
BUILDING MODELS FOR ANALYZING SYSTEMS
• Systems analysis is the process of investigating the behavior of the system by choosing
various options for control. It is an organized study of alternatives and options for the design
and management of a system (natural or engineered). We create mathematical models to assist
us in the design and management of such natural or engineering systems.

• It is a coordinated set of procedures that can be used to address issues of engineering design,
project planning and management.

• The models also assist us in developing policy for the systems. They consider both technological
and economic factors. Principles of engineering are combined with principles of economics to
achieve optimal solutions.

TYPES OF MODELS
• Iconic Models: These are physically built prototypes such as paper airplanes, model cities and
villages, model cars etc.

• Participatory Models: Role plays, monopoly, chess, war games, fire drills, flight simulators etc.

• Mathematical Models: A set of mathematical equations used to describe the system. For
example, finite element models describe the behavior of structures; kinematic wave equations
describe the flow of water; econometric models describe the current economic conditions.
Mathematical models are built using differential equations to explain and predict natural
phenomena given some initial conditions. They are called descriptive models. For a given set
of input data and initial conditions, these models offer a description of the outputs of the system
through time. For instance, species growth models predict the growth of a species given its initial
population and growth parameters. The advent of modern computers has made it possible to
develop global models that mimic natural phenomena. The global climate models, for instance,
are a set of physics based equations that simulate the climate using some initial conditions.
Since World War II, the focus of mathematical modeling has shifted towards the science of
decision-making and policy development. This formed the evolution of prescriptive models
that prescribe a course of action, a design or a policy using the mathematics of decision-making.

Descriptive Models Prescriptive Models


If I follow this course of action, what will hap- What is the best course of action to follow?
pen.
Predicts the outcome given initial conditions What is the best strategy among all possible
outcomes?

Prescriptive models have cost, and are driven by the motive to maximize or minimize an
objective. They allocate resources (broadly classified into labor, power, money, and material) in
an effective manner. They are also called optimization models due to their best outcome
(optimal) feature. Descriptive models are often a part of prescriptive models in simulations.

Lecture Notes Page 1 Devineni


• Deterministic and Stochastic Models: Models can also be classified based on the type of
data they use. If the data and outcomes are relatively certain, i.e., the parameter values are
known with certainty, these models are referred as deterministic. For example, the number of
stock piles in a factory can be determined using data on initial stock plus the new stock
purchased. The number of wooden pieces of a specific size that can be cut from a big plywood
are fixed - deterministic.
If the data inputs and parameters are random variables, the models are stochastic. For
example, water storage in a reservoir is equal to initial storage + the input that depends on the
probability of rainfall. Hence, storage is described as a random variable with a mean and
variance.

Figure 1: Deterministic and Stochastic processes

Deterministic Stochastic
Prescriptive Linear Programming (widely used) Stochastic Programming
Integer Programming (parameters are random variables)
Multiobjective Programming
Dynamic Programming
Descriptive Differential Equations Stochastic differential equations (dif-
(Applied math and calculus using de- ferential equations + random param-
rived parameters and rate constants, eters)
initial conditions) Queueing theory
Can be linear or non-linear Monte Carlo simulations that gener-
ate realistic events and system re-
sponses

Lecture Notes Page 2 Devineni


• Statistical Models: A statistical model explains the stochastic process with some certainty
(explained variability) and some error (unexplained variability). It is a hypothesis of the
relationship between outputs and inputs that tests and provides expected outcomes given some
inputs.

SYSTEMS ANALYSIS OR OPERATIONS RESEARCH


As the complexity and specialization in an organization increases, it becomes more and more difficult
to allocate the available resources to the various activities in a way that is most effective for the
organization as a whole. The need to find better solutions to these problems provided the environment
for the emergence of systems analysis or operations research (OR). The origins for OR can be
traced back to early World War II. There was an urgent need to allocate scarce military resources to
operations in an effective manner. They started doing “research” on military “operations.”

The nature of OR:

• Careful observation and formulation of the problems and gathering relevant data.

• Construct mathematical models with reasonable representation of the real-world constraints


(financial, physical, and institutional constraints). Financial constraints are generally the limited
supply of money and the cost of borrowing it. This money may be used to obtain resources
(material or hiring workers). Physical constraints refer to the limitations of the properties of
materials (e.g., strength, elasticity, engineering properties). Institutional constraints are generally
rules, laws or guidelines specified by society, government, and the engineering profession. For
example, design and building codes are institutional constraints.

• Test the hypothesis and validate the models.

• Objectives must be consistent with the overall goals of the company. Profit-maximization,
cost-minimization, societal benefit etc.

• Search for optimality.

• Provide understandable conclusions to decision makers.

• It is a team effort with experts in Engineering, Mathematics, Statistics, Probability, Economics,


Business, Computer Science, etc.

OVERVIEW OF THE SYSTEMS/OR MODELING APPROACH


The following are the several phases involved in systems modeling and analysis.

Defining the problem and gathering data


Study the system and develop a well defined statement of the problem. Determine appropriate
objectives, constraints, inter-relations between other organizations, alternative course of actions and
time limitations. Perform detailed technical analysis and provide recommendations. Offer alternative
attractive options that need benefit cost analysis. The objectives should be formulated to benefit the
overall organization. The final decision lies with the management.

Lecture Notes Page 3 Devineni


Long run profit maximization is a good organizational objective. This considers the time value of
money.

The OR team needs a good handle on data sources. Often, much of the data may be “soft,” i.e.,
rough estimates based on educated guesses.

Formulating a mathematical model


Mathematical models are idealized representations of real systems expressed as symbols and equations.
The terminology in a standard OR problem are:

Decision Variables: x1 , x2 , · · · xn are the values for the quantifiable decisions to be made.

Objective Function: The mathematical function of profit or cost. Z = c1 x1 + c2 x2 + · · · + cn xn

Constraints: Restrictions on the values that can be assigned to the decision variables expressed by
equalities or inequalities.

Parameters: The constants in the objective function and the constraints.

Determining the parameters require careful study of the problem and gathering required data. These
values are mostly rough estimates for most real-world problems. Since there is uncertainty in these
values, it is important to perform sensitivity analysis that informs us as to how the solution will
change with changing parameters. One can develop multiple models to represent the same system.
Each model answers specific questions.

Because mathematical models are idealized representations of the real-world, we have to make
simplified assumptions to make the model tractable. Model testing and validation stages help in
improving the model construction. There is always a trade-off between precision of the model and the
tractability of the model.

Deriving solutions from the model


A common theme in OR is the search for optimality or best solution. If the model is well developed
and tested, the results of the optimality are a reasonable approximation of what can be implemented
in the real-world.

The solution has to be Satisficing = Satisfactory + Optimizing , a term first coined by Herbert
Simon (Nobel Prize winner in economics).

Establish minimum satisfactory levels of performance in various areas based on past performance or on
what the competition is achieving. “Optimizing is the science of ultimate, Satisficing is the art of
feasible.”

We should also focus on Post Optimality analysis that involve developing what-if scenarios to test
the outcomes under different conditions and changes. In other words, a sensitivity analysis is necessary

Lecture Notes Page 4 Devineni


to check which parameters are most critical. Sensitive parameters are the parameters whose values
cannot be changed without changing the optimal solution.

Testing the model


The process of testing and improving the model and increasing its validity is called model validation.
Look at the overall model after formulating. Check for units of the parameters. Re-examine the
definition of the problem. Perform retrospective tests using historical data.

Preparing to apply the model


Have a well documented system for applying the model. Create a decision support system to help
managers use the data and models to support their decision-making process.

Implementation
A phase where all the benefits will be reaped. It needs the consent of senior management. They need
to be informed from the outset. Provide a careful explanation of the new system to be adopted.
Develop the procedure required for implementation. Obtain feedback on the performance and
assumptions.

Rules for Modeling


• Keep the model simple. Use only parameters and data that is readily measurable or with low
uncertainty.

• Let x = unknown level of decision. Where multiple decisions are involved, let xi = unknown level
of decision for i.

• List all the constraints and articulate the objectives. Objective is the goal of the problem. All
constraints have to be met.

• Try to develop linear or additive models for simpler solutions and tractability.

General Caution
You can use the solutions for decision-making.
The solution itself is not hard bound optima.
The other outputs from the post-optimality
analysis are more useful in understanding the
system and developing or designing measures to
manage the problem.

SOME APPLICATIONS
1. Distribution, Warehousing, and Industrial Siting: Least cost solution for the distribution of goods
from multiple locations to multiple destinations. These are typically called routing problems.

2. Solid Waste Management: Siting landfills, collection points, routing trucks.

3. Manufacturing, Refining and Processing: Factory decisions of what items to manufacture.

Lecture Notes Page 5 Devineni


4. Educational Systems: Class scheduling, room scheduling, school bus routing, managing
enrollments.
5. Personnel Scheduling and Assignment: Shift rotation in factory, assign jobs in large
organizations, scheduling airline crew, scheduling snow removal on highways.
6. Emergency Systems: Siting of fire engines, ambulances, trucks etc., to minimize travel time.
7. Transportation Sector: Design transportation networks with efficiency and connectivity as
criteria. Bus stop locations, traffic light timing, entrance ramps.
8. Sales: Traveling salesman has to cover districts with equal sales potential.
9. Electric Utilities and Air Quality Management: Power plant siting and installing transmission
lines.
10. Tele-communications: Design of communication networks.
11. Water Management: Design and operation of reservoirs, water distribution networks, sewer
networks, water quality management.
12. Agriculture: Crop choices, water and fertilizer management.
13. Civil Infrastructure: Cut and fill, alignment of highways, construction scheduling.

EXAMPLES
1. Siting Health Clinics
To site 10 health clinics in a region. Communities with 1000 or more people should be able to
reach a clinic within 10 miles. Clinic is only sited in a community with 10,000 or more people.
Data: Population of each community, distance between the communi-
ties.

Objective: Average distance traveled by all users is minimized.

Constraints: Only communities with 10,000 or more people can have a


clinic, There are only 10 clinics.

Decision Variables: Which communities are assigned to which clinics


and where.
2. Building Reservoir for Water Supply
Data: Projected monthly demand for the planned period, worst drought
in the history, historical streamflow.

Objective: min(Smax )

Constraints: S ≤ Smax ; S ≥ Smonth

Decision Variables: Monthly storage S and maximum storage capacity


Smax .

Lecture Notes Page 6 Devineni


3. Upgrading the links in a highway network
Data: Population, cost of construction, projected trade.

Objective: Maximize connectivity or maximize potential trade between


cities.

Constraints: Budget.

Decision Variables: Which cities to connect.


4. Siting fire stations in a city
Divide the city into zones. Each zone should have an eligible fire station that is within a
5-minute driving radius. Fire stations can be placed anywhere. The goal is to have the least
number of stations while meeting all demands.

5. Expansion of Electric Power System


New power lines will be sited near the grids. There is projected demand for energy. Siting has to
be done at low cost of building and maintenance.

Lecture Notes Page 7 Devineni


MODEL FORMULATION - EXAMPLES FROM CIVIL ENGINEERING
All branches of civil engineering develop and implement optimization models to guide decision-making
and policy development. Some example formulations are simplified and presented here. These models
can get more complex in real-world decision-making.

There is an underlying common structure of model formulation that represents the fundamental way of
viewing engineering decision problems. It contains:
• Cost related objective function that need to be minimized, or a desired objective that need to be
maximized.
• Constraints limit the range of possible activities.
• Models can be linear or non-linear on the objective functions and constraints.
For example, consider an activity analysis problem where we need to determine how many items of a
given type should be manufactured. Each item brings in a certain profit, and it increases as the
production level increases. Each item utilizes some resources that are limited. Let us use the following
notations to formulate the model:

i = 1 · · · n Þ number of decision variables

j = 1 · · · m Þ number of constraints

xi Þ quantity or extent of variable i

ci Þ unit profit of quantity i

ai j Þ coefficient of ith quantity of jth constraint. It represents the amount of resource j that i
consumes.

b j Þ available resource j

Objective: Maximize profit,

Z = c1 x1 + c2 x2 + · · · + cn xn

subject to Constraints:

a j1 x1 + a j2 x2 + · · · + a jn xn , ∀ j = 1···m

Non-negative Constraints:

xi ≥ 0 ∀i = 1 · · · n
We can either maximize or minimize the objective function Z, depending of whether it is profit or cost.
All the variables are on the left hand side. All the constants are on the right hand side. The signs are
≤, ≥ or =

Lecture Notes Page 8 Devineni


Often, Z may not be linear. In this case, you can explore transformations to make it linear. Objective
functions that cannot be linearized have to be solved using other non-linear methods.

Non-linear model formulation


Objective: Maximize profit, Z = f (x1 , x2 , · · · , xn ) where f is the non-
linear function.

subject to Constraints: g(x1 , x2 , · · · , xn ) ∗ b j ∀ j = 1 · · · m where


∗ is ≤, ≥ or = and g is the non-linear constraint function.

Blending Water Supplies Example


A city needs new water supply for its growing population. The department of water supply has
identified new aquifers for this purpose but need blending with old sources to reduce hardness. The
final blended water should not have a hardness more than 1200 lbs per million gallons. A
minimum of 150 additional million gallons per day of water is needed. The management is
interested in a least-cost strategy for expanding water supply while ensuring good water quality.

They have three sources, source 1 Þ current supply, source 2 Þ new aquifer, and source 3 Þ distant
stream. The cost to obtain water in $ per million gallons, supply limits in mgd, and hardness in lbs per
million gallons are:

S1 S2 S3
Cost ($/mgd) 500 1000 2000
Supply Limit (mgd) 25 120 100
Hardness (lbs/mg) 200 2300 700
Decision Variables
x1 = mgd from S1
x2 = mgd from S2
x3 = mgd from S3

Objective Function: Minimize Total Cost


Z = 500x1 + 1000x2 + 2000x3

Constraints:
Water supply constraint: x1 + x2 + x3 ≥ 150 mgd

Hardness constraint: 200x1 + 2300x2 + 700x3 ≤ 1200 ∗ (x1 + x2 + x3 )

Sustainability constraint: x1 ≤ 25 mgd (Source 1)


Sustainability constraint: x2 ≤ 120 mgd (Source 2)
Sustainability constraint: x3 ≤ 100 mgd (Source 3)

Non-negative constraint: x1 ≥ 0; x2 ≥ 0; x3 ≥ 0

Lecture Notes Page 9 Devineni


Furniture Factory Example
A furniture factory specializes in making furniture for dormitories. They make student desks, desk
chairs, dressers and tables and use maple wood for exterior surfaces and pine wood for interior
surfaces. Assume other resources such as shop and inventory space, and labor are adequate, what
product should they make to maximize profit?

Item Desks Chairs Dressers Tables Amount Cost


Available ($/board-
(board-feet) feet)
Pine 7 1.5 10 4 4000 2
Maple 15 4 22 10 2000 4
Revenue $150 $100 $250 $170
The above table provides information on the manufacturing process. Each chair that the company
produces requires 1.5 board-ft of pine and 4 board-ft of maple. It sells for $100. A total of 4000
board-ft of pine and 2000 board-ft of maple is available at a cost of $2 per board-ft for pine and $4
per board-ft for maple.

Decision Variables
x1 = # of desks
x2 = # of chairs
x3 = # of dressers
x4 = # of tables

Objective Function: Maximize Profit


Z = 150x1 − (2)(7)x1 − (4)(15)x1
+ 100x2 − (2)(1.5)x2 − (4)(4)x2
+ 250x3 − (2)(10)x3 − (4)(22)x3
+ 170x4 − (2)(4)x3 − (4)(10)x4

Z = 76x1 + 81x2 + 142x3 + 122x4

Constraints:
Pine Resource: 7x1 + 1.5x2 + 10x3 + 4x4 ≤ 4000 board-ft

Maple Resource: 15x1 + 4x2 + 22x3 + 10x4 ≤ 2000 board-ft

Non-negative constraint: x1 ≥ 0; x2 ≥ 0; x3 ≥ 0; x4 ≥ 0

Lecture Notes Page 10 Devineni


Grading a Portion of a Highway
Highway through a hilly terrain is to be graded to elevations that meet acceptable standards of vertical
curvature. Slope and rate of change of slope of the road must be within preset bounds. Assume two
areas are to be cut and three areas are to be filled.

Figure 2: Cut and Fill Problem

You cannot cut more than a1 from cut portion 1 and a2 from cut portion 2. You have to fill atleast b1 ,
b2 , b3 in fill portions 1, 2 and 3 for the designed slope. The cost of transporting cut portion i to fill
portion j is ci j .

Decision Variables
xi j where i = 1, 2 and j = 1, 2, 3 is the amount cut from i and transported to fill j

Objective Function: Minimize Total Cost


2 3
Z = ∑ ∑ ci j ∗ xi j
i=1 j=1

Constraints:
Cut Constraint: x11 + x12 + x13 ≤ a1

Cut Constraint: x21 + x22 + x23 ≤ a2

Fill Constraint: x11 + x21 ≥ b1

Fill Constraint: x12 + x22 ≥ b2

Fill Constraint: x13 + x23 ≥ b3

Non-negative constraint: xi j ≥ 0; ∀i, j

Lecture Notes Page 11 Devineni


Design a Building for Cost - Non-linear Formulation
A new square shaped office building has to be constructed to provide 400,000 f t 2 area

Decision Variables
x = length in feet of the side
z = number of floors

Objective Function: Minimize Total Cost = Foundation cost + Building


cost
Z = Ax2 z1/4 + Bx2 z3/2

Constraints:
Area Constraint: x2 z ≥ 400, 000 f t 2

Non-negative constraint: x ≥ 0; z ≥ 0

Cleaning the River


Two communities C1 and C2 are discharging untreated waste (high in organic content) into the river.
Oxidation of organic matter by microbes results in low dissolved oxygen (DO), bad odor and depletion
of fish population. For good stream health, the DO should be maintained at a specified level.
Both communities agreed to reduce organic matter in their waste by certain quantity before they
discharge. They will share the cost of clean up.

Figure 3: Water Quality

The DO at C2 is a function of the organic matter removed at C1 . Similarly, the DO below C2 is a


function of organic matter removed at C1 and C2 . These relations are given by:

Lecture Notes Page 12 Devineni


1 + 0.1ε1 = DO at C2

2 + 0.02ε1 + 0.02ε2 = DO below C2

ε1 and ε2 are the organic removal efficiency (%) at the treatment plants. Both efficiency levels have to
be at least 35%. Further, the DO at C2 should be at least 6 mg/l and the DO below C2 should be at
least 4 mg/l. The cost of removal for community is $20,000 per efficiency fraction and $10,000 per
efficiency fraction for communities 1 and 2 respectively.

Decision Variables
ε1 = removal efficiency for community 1
ε2 = removal efficiency for community 2

Objective Function: Minimize Total Cost


Z = 20000ε1 + 10000ε2

Constraints:
Water Quality Constraint at C2 : 1 + 0.1ε1 ≥ 6 mg/l

Water Quality Constraint below C2 : 2 + 0.02ε1 + 0.02ε2 ≥ 4 mg/l

Minimum Removal Constraint at C1 : ε1 ≥ 35%

Minimum Removal Constraint at C2 : ε2 ≥ 35%

Non-negative constraint: ε1 ≥ 0; ε2 ≥ 0

Lecture Notes Page 13 Devineni


Selecting Projects for Bidding - Zero-One Binary Integer Problem
A contractor has to apply for bids out of eight projects. He cannot undertake all the eight projects due
to limited resources. Hence, he has to make a choice on what projects to bid for. Details are given in
the following table.

Project Time to Prepare (Hrs) Hours of skilled labor Cranes required? Profit (in $1000)
1 8 6000 1 80
2 12 5000 0 110
3 13 4000 1 100
4 11 7000 0 90
5 9 8000 0 70
6 7 3000 1 80
7 8 4000 1 90
8 8 5000 1 60
The contractor has 3 cranes, 30,000 hours of skilled labor, and a total of 50 person-hours to prepare
the bid. Given this information, which of the eight projects should be pursued?

Decision Variables
x j ∀ j = 1 : 8 = 1 or 0; 1 = yes, go for the bid, 0 = no

Objective Function: Maximize Profit


Z = 80x1 + 110x2 + 100x3 + 90x4 + 70x5 + 80x6 + 90x7 + 60x8

Constraints:
Cranes Constraint : 1x1 + 0x2 + 1x3 + 0x4 + 0x5 + 1x6 + 1x7 + 1x8 ≤ 3 cranes

Person Hours Constraint: 8x1 + 12x2 + 13x3 + 11x4 + 9x5 + 7x6 + 8x7 + 8x8 ≤ 50 person-hours

Workforce Constraint: 6x1 + 5x2 + 4x3 + 7x4 + 8x5 + 3x6 + 4x7 + 5x8 ≤ 30 hours

Integer Constraint: x j ∈ (0, 1)∀ j

Lecture Notes Page 14 Devineni


Tent Design
Imagine you are in a scout camp. You are given a tarpaulin that is 10 feet by 10 feet and are asked to
construct a vertically symmetric tent with straight sides that are open on both ends. A single rope
stretched between two tent poles will support the tent. The sides of the tent will be anchored with
tent stakes. How will you design this tent?

Volume of the tent V = 12 bh ∗ (10) has to be maximized.

Decision Variables
b, h base and height of the tent

Objective Function: Maximize Volume


Z = 5bh

Constraints:
Geometry Constraint: ( b2 )2 + h2 = 52 Pythagorus Theorem for the right triangle

Non-negative constraint: b ≥ 0; h ≥ 0

Air Pollution Control


Two coal plants (distantly located) provide electricity for two major cities. Due to high altitude winds,
the emissions from the plants’ tall stacks get transported downwind to the cities. There is a decision
to reduce SO2 from the plants. Currently, the plants discharge 12,000 and 20,000 kilograms of SO2
per month. They should be cut back in a way that reduces the sulfate deposition at cites 1 and 2 to
less than 50 kg. The cost of reduction is $1000 per kg. The transport coefficients are as follows:

City 1 City 2
Plant 1 0.0075 0.0025
Plant 2 0.0025 0.0075
Decision Variables
x1 = removal at plant 1
x2 = removal at plant 2

Objective Function: Minimize Total Cost


Z = 1000x1 + 1000x2

Constraints:
Air Quality Constraint at city 1: (12000 − x1 ) ∗ 0.0075 + (20000 − x2 ) ∗ 0.0025 ≤ 50 kg

Air Quality Constraint at city 2: (12000 − x1 ) ∗ 0.0025 + (20000 − x2 ) ∗ 0.0075 ≤ 50 kg

Non-negative constraint: x1 ≥ 0; x2 ≥ 0

Lecture Notes Page 15 Devineni


Land and Species Preservation
An agricultural land is being converted for development. They will put in species preservation for some
parcels. You have to come up with the smallest number of parcels that contain most species.
Following are the details on which parcel has what species.

Parcel Species
1 A, B
2 D
3 A, C
4 C
5 A, D
6 B
7 B
8 B
9 C, D
Decision Variables
x j ∀ j = 1 : 9 = 1 or 0; 1 = yes, preserve, 0 = no

Objective Function: Minimize Parcels


Z = x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8 + x9

Constraints:
Species Constraint : x1 + x3 + x5 ≥ 1 species A
Species Constraint : x1 + x6 + x7 + x8 ≥ 1 species B
Species Constraint : x3 + x4 + x9 ≥ 1 species C
Species Constraint : x2 + x5 + x9 ≥ 1 species D

Integer Constraint: x j ∈ (0, 1)∀ j

Lecture Notes Page 16 Devineni


STUDY GUIDE
Goal
Learning the basic concepts of systems analysis and model formulation.

Objectives
1. To develop the ability to identifying a system(s) and its various components and controls.

2. To develop the ability to formulate an optimization model by identifying decision variables,


objective function(s), and constraints.

Tasks
1. Read the following documents in the listed order.

(a) Lecture Notes 1 - Introduction to Civil Engineering Systems Analysis.


(b) Textbook Chapter 1 and Chapter 2.
(c) The article “yellowfreight.pdf.” Write a one-page summary of the application and the
benefits the project provided.
(d) The article “netherlands waterplanning.pdf.” Write a one-page summary of the application
and the benefits the project provided.
(e) The article “hall 1985.pdf.” Summarize the author’s viewpoint about the complementary
roles of modeling, evaluating information from the model, and then applying the decision
maker’s judgment when deciding on a course of action.
(f) The article “miser 1989.pdf.” Summarize the author’s viewpoint about the roles of
observation and experimentation in the model validation process.
(g) The article “gass 1983.pdf”. What does the author say about whether a model can be
completely validated. Describe the role of sensitivity analysis in testing the operational
validity of a model.

2. Solve the following problems to understand the concepts. Summarize the key concepts learned
from them.

(a) Chapter 1 - Exercises 1.1 to 1.6, and 1.8.


(b) Chapter 2 - Exercises 2.4, 2.5, and 2.7.

3. Timed Workouts: Simulate an exam-like environment and solve the problems given in the
“timed workouts” section under the specified time. This will help you prepare for modular
assessments and exams.

Lecture Notes Page 17 Devineni


TIMED WORKOUTS
Tom’s Timber Company (15 minutes time limit)
Tom’s Timber Company produces precision cut timber at its two factories.

Albert, Wilbert, and Gilbert placed orders for 400 units, 300 units and 300 units of timber respectively
from Tom.

Factory 1 can produce at most 500 units and Factory 2 can produce at most 500 units.

Cost of shipping from the two factories to the three customers are as follows:

• $600 → Factory 1 to Albert; $400 → Factory 2 to Albert

• $800 → Factory 1 to Gilbert; $900 → Factory 2 to Gilbert

• $700 → Factory 1 to Wilbert; $600 → Factory 2 to Wilbert

Tom should determine how many units of timber to ship from each factory to each of their three
customers. He has to come up with an optimal plan to send the lumber supplies at the lowest cost.

1. In words, identify the decision variables, objective function, constraints, and the model
parameters for Tom’s optimal plan.

2. The mathematical description of Tom’s optimal plan is given below.


n m
Z = ∑ ∑ ci j ∗ xi j
i=1 j=1

m
∑ xi j ≤ ai ∀i = 1 : n
j=1
n
∑ xi j ≥ b j ∀j = 1 : m
i=1

From the information given about the problem, write down what each of these math symbols
describe: Z, n, m, ci j , xi j , ai , b j

3. Should Tom maximize or minimize his objective function?

4. There is an important constraint missing. Identify what that is and describe it mathematically.

Lecture Notes Page 18 Devineni

You might also like