HO CHI MINH CITY INTERNATIONAL UNIVERSITY

SCHOOL OF INDUSTRIAL ENGINEERING & MANAGEMENT

PROJECT REPORT

TOPIC: PROFIT OPTIMIZATION MATH PROBLEM

COURSE: MULTI-CRITERIA DECISION MAKING

LECTURER: Dr. Đào Vũ Trường Sơn

Group members:

Nguyễn Phi Long MIEIU22003

Nguyễn Kiều Việt Hưng MIEIU22002

Trần Duy Anh MIEIU22001

 Table of contents
I. INTRODUCTION 2

1. About MCDM: 2

2. Problem objectives: 3

3. Problem scope: 3

4. Problem limitation: 3

II. CASE STUDY 4

III. METHODOLOGY 5

1. Mathematical model: 5

2. Applying Excel Solver: 6

3. Applying Lingo: 9

IV. SENSITIVE ANALYSIS 13

V. CONCLUSION 13

VI. REFERENCES 13

1
I. INTRODUCTION

1. About MCDM:

Multi-criteria decision making (MCDM) is a branch of operations research. Decision making often
involves imprecision and vagueness which can be effectively handled by fuzzy sets and fuzzy
decision making techniques. In recent years, a great deal of research has been carried out on the
theoretical and application aspects of MCDM and fuzzy MCDM. Methods of multicriteria
decision-making (MCDM) can be used for breaking down complex problems into manageable
components. With the help of MCDM, different dimensions that are important for the
decisionmaking context may be considered and evaluated one at a time. With the help of group
decisionmaking methods, opinions from several decision-makers (possibly having different values
and preferences) can be collected and included in the decision. In general, the MCDM process
consists of several steps that include:

• Definition of the overall objective for the decision-making.

• Dividing the goal to several lower-level objectives that describe different dimensions
relevant for reaching the goal.
• Defining the criteria that describe the performance of the alternatives in each selected
dimension.
• Defining the decision alternatives and collecting data on their performance and
characteristics.
• Creating a matrix that describes the performance of the alternatives on each selected
criteria.
• The actual decision-making, consisting of criteria weightings by the participating
decisionmakers and ranking of the decision alternatives.

MCDM includes MODM and MADM:

• MODM (Multiple Objective DM): Finding out the set of non-dominated solutions from
infinite solutions, according to multiple conflicting objectives.
• MADM (Multiple Attribute DM): selection of the best solution from a set of alternatives,
each of which is evaluated against multiple conflicting criteria.

2
2. Problem objectives:

This project will focus on how MADM/MODM methods can be applied practically. Moreover,
this opportunity is the masterpiece chance to apply modern method to solve complex problem
involving management of large systems, as well as empower decision making in problem
administration. In this project, our team will solve this problem by using Excel’s SOLVER, an
application of Excel that helps find optimal result for small scale linear problem and LINGO, a
mathematic solving program. By developing a mathematical model and assumed data, we will
present a feasible amount of each blend for the coffee shop which make owner have the maximum
profits.

3. Problem scope:

The scope of Operation research is to design a model to convert business problems into
mathematical hypotheses, while also specifying expected behavior and goals, using modern
methods such as Excel's SOLVER and Lingo. Not only in Industrial Engineering and
Management, but also in a wide range of other sectors, from material procurement to the
dispatch of finished products. Operation Research approach helps in operation
management. The operating systems convert the inputs to the satisfaction not only the
customers need but also the profits of the owners.

4. Problem limitation:
Due to time constraints and a lack of data and program in this progress, our Excel Solver
and LINGO results may not be the best answer for this project and may be time-consuming
influence on the decision of the owner. Therefore, if any businesses want to use our solution
in their stores, they must be careful to understand the practical results for each type of
coffee, such as market price, validity, or the actual proportion of mix required... However,
when statistics vary, other methods should be applied into consideration.

3
II. CASE STUDY

The Urban Coffee Shop blends coffee on the premises for its customers. It sells three basic blends
in 1-pound bags, Special (s), Mountain Dark (d) and Mill Regular (r). It uses four different types
of coffee to produce the blends, Brazilian (b), Mocha (o), Columbian (c) and Mild (m).

- Special (s): At least 35% Columbian (c), at least 30% Mocha (o). Profit: $6.50 per 1-pound bags.

- Mountain Dark (d): At least 55% Brazilian (b), no more than 10% Mild (m). Profit: $5.25 per
1pound bags.

- Mill Regular (r): No more than 60% Mild (m), at least 25% Brazilian (b). Profit: $4 per 1-pound
bags.

- The cost of Brazilian (b) coffee: $2.50 per pound.

- The cost of Mocha (o) coffee: $3.25 per pound.

- The cost of Columbian (c) coffee: $3.30 per pound.

- The cost of Mild (m) coffee: $2.20 per pound.

The shop has 120 pounds of Brazilian coffee, 80 pounds of Mocha, 70 pounds of Colombian, and
140 pounds of Mild coffee available per week.

The shop wants to know the amount of each blend it should prepare each week in order to maximize
profit. Formulate and solve a linear programming model for this problem.

4
III. METHODOLOGY

1. Mathematical model:

- Decision variables:

Let

𝑥𝑏𝑠, 𝑥𝑏𝑑, 𝑥𝑏𝑟: the quantity (in pounds) of Brazilian coffee to be used to make Special, Dark and
Regular blend respectively.

𝑥𝑜𝑠, 𝑥𝑜𝑑, 𝑥𝑜𝑟: the quantity (in pounds) of Mocha coffee to be used to make Special, Dark and Regular
blend respectively.

𝑥𝑐𝑠, 𝑥𝑐𝑑, 𝑥𝑐𝑟: the quantity (in pounds) of Colombian coffee to be used to make Special, Dark and
Regular blend respectively.

𝑥𝑚𝑠, 𝑥𝑚𝑑, 𝑥𝑚𝑟: the quantity (in pounds) of Mild coffee to be used to make Special, Dark and Regular
blend respectively.

- Objective function:

Max Z = (6.50 – 2.50)𝑥𝑏𝑠 + (5.25 – 2.50)𝑥𝑏𝑑 + (4 – 2.50)𝑥𝑏𝑟 + (6.50 – 3.25)𝑥𝑜𝑠 + (5.25 – 3.25)𝑥𝑜𝑑
+ (4 – 3.25)𝑥𝑜𝑟 + (6.50 – 3.30)𝑥𝑐𝑠 + (5.25 – 3.30)𝑥𝑐𝑑 + (4 – 3.30)𝑥𝑐𝑟 + (6.50 – 2.20)𝑥𝑚𝑠 + (5.25 –

2.20)𝑥𝑚𝑑 + (4 – 2.20)𝑥𝑚𝑟

 Max Z = 4𝑥𝑏𝑠 + 2.75𝑥𝑏𝑑 + 1.5𝑥𝑏𝑟 + 3.25𝑥𝑜𝑠 + 2𝑥𝑜𝑑 + 0.75𝑥𝑜𝑟 + 3.2𝑥𝑐𝑠 + 1.95𝑥𝑐𝑑 + 0.7𝑥𝑐𝑟 + 4.3𝑥𝑚𝑠

+ 3.05𝑥𝑚𝑑 + 1.8𝑥𝑚𝑟

- Constraints:

𝑥𝑏𝑠 + + ≤ 120

𝑥𝑜𝑠 + + ≤ 80

𝑥𝑐𝑠 + + ≤ 70

𝑥𝑚𝑠 + 𝑥𝑚𝑑 + 𝑥𝑚𝑟 ≤ 140

5
𝑥𝑐𝑠 ≥ 0.35( + + + )

𝑥𝑜𝑠 ≥ 0.3( + + + )

𝑥𝑏𝑑 ≥ 0.55( + + + )

𝑥𝑚𝑑 ≤ 0.1( + + + )

𝑥𝑚𝑟 ≤ 0.6( + + + )

𝑥𝑏𝑟 ≥ 0.25( + + + )

𝑥𝑏𝑠, 𝑥𝑏𝑑, 𝑥𝑏𝑟, 𝑥𝑜𝑠, 𝑥𝑜𝑑, 𝑥𝑜𝑟, 𝑥𝑐𝑠, 𝑥𝑐𝑑, 𝑥𝑐𝑟, 𝑥𝑚𝑠, 𝑥𝑚𝑑, 𝑥𝑚𝑟 ≥ 0

2. Applying Excel Solver:

First, we need to set up the formula of the problem as the image below. Next, we choose Data >>
Solver to set up the formula to solve the problem.

We set $G$32 like the objective function we need to find

The changing variable cell $G$7:$G$9 as variables 𝑥𝑏𝑠, 𝑥𝑏𝑑, 𝑥𝑏𝑟, 𝑥𝑜𝑠, 𝑥𝑜𝑑, 𝑥𝑜𝑟, 𝑥𝑐𝑠, 𝑥𝑐𝑑,

𝑥𝑐𝑟, 𝑥𝑚𝑠, 𝑥𝑚𝑑, 𝑥𝑚𝑟.

6
And all constrains of problem in Subject to the Constraints. Finally, we use button Solve to find a
solution.

7
After solving, we have:

So, the maximum profit for the coffee shop is $1181 and the amount of each blend it should prepare
each week are:

• The amount of Brazilian (b) coffee: 120 pounds.

• The amount of Mocha (o) coffee: 80 pounds.

• The amount of Columbian (c) coffee: 70 pounds.

• The amount of Mild (m) coffee: 140 pounds.

8
3. Applying Lingo:

Firstly, we input the given data to Lingo:

Then, we click on the bulleyes icon named ‘Solve’ to solve the model. It then gives the results
below:

9
Thus, the table shows how many bags for each blend the shop should order. Also, the maximum
profit of the Urban Coffee shop is $1181.

Comparing to the result of Excel Solver, we can see that it is the same with Lingo’s.

10
IV. SENSITIVE ANALYSIS

Before we start, we need to set an option in LINGO. After that, we’ll proceed with a simple
example.

Setting Options for Ranges in LINGO

There is a particular tab that needs to be changed in

order to get a range report in LINGO. Under Solver, go
to Options. Find a tab at the top of the dialog box that
says General Solver, then find the box titled Dual
Computations. Set that box so compute Prices and
Ranges (the default is prices only). You should select
Apply and Save.

To see the range for each variable, we have to click (or select) the model window, then select
Range under the Solver tab as shown below, which should result in the range report (right image,
below).

Ranges in which the basis is unchanged:

Objective Coefficient Ranges:

Current Allowable Allowable
Variable Coefficient Increase Decrease
XBS 4.500000 3.000000 INFINITY
XOS 3.750000 3.000000 6.166667
XCS 3.600000 INFINITY 4.625000
XMS 4.800000 INFINITY 3.000000
XBD 3.250000 0.000000 1.666667
XOD 2.500000 6.166667 0.000000
XCD 2.350000 4.625000 INFINITY
XMD 3.550000 10.62500 2.500000
XBR 1.750000 1.666667 0.000000
11
 XOR 1.000000 0.000000 INFINITY
XCR 0.8500000 4.625000 INFINITY
XMR 2.050000 2.500000 0.7870370

Right-hand Side Ranges:

Current Allowable Allowable

Row RHS Increase Decrease
2 0.000000 12.88889 13.33333
3 0.000000 10.00000 29.00000
4 0.000000 11.60000 INFINITY
5 0.000000 29.00000 6.000000
6 0.000000 11.50000 19.33333
7 0.000000 13.80000 INFINITY
8 110.0000 690.0000 32.22222
9 70.00000 18.12500 10.00000
10 80.00000 13.33333 17.57576
11 150.0000 48.33333 76.66667

Note that these are the allowable changes, not the actual values, by which we can change the
current values (of the coefficients shown or the RHS of the constraints) and maintain the current
set of basic variables.

For example, the 4.5 for XBS can be increased to 7.5, and we would still have the same set of basic
variables comprise the BFS at which the optimum occurs.

By the way, we assume these changes are done one at a time. There is an analysis one can perform
for multiple changes, but we will not look at that this semester.

12
V. CONCLUSION

In conclusion, this project has brought a lot of benefits and valuable experiences to our group. We
had the opportunity to review the knowledge and skills learned in MCDM classes during the last
semester. We were able to research and study the data of the stores to set up the problem for this
project. Through research and discussion, we have selected tools to help solve the problem and
complete the project. And finally, we solved the problem and made comments on the problem, the
results and the project. This project was a great opportunity for us to meet, work, research together,
apply our knowledge, give personal perspectives to complete the project in the best way.

VI. REFERENCES

David Poole. (2011). Linear Algebra: A Modern Introduction 3rd Edition. Thomson Brooks/Cole

https://www.sciencedirect.com/topics/engineering/multi-criteria-decision-making

https://www.collectors2020.eu/tools/multi-criteria-decision-making/

https://www.mbacrystalball.com/blog/operations-management/operations-research/

13