Journal of Applied and Fundamental Sciences

DECISION MAKING WITH SIMPLEX MODEL: A

CASE STUDY
Bapan Kalita1*, Abdul Barik Sarker2, Sankar Barman3
1,2
Department of Mathematics, Assam Royal Global University, Guwahati, Assam, India
3
Department of Physics, Assam Royal Global University, Guwahati, Assam, India
*For correspondence. ([email protected])

Abstract: Decision making is a big deal for a manufacturer. Simplex model is a vibrant tool for the decision
makers. This paper focuses on of the most used simplex technique to take the decision of producing optimum
number of items for earning maximum profit. Here, a drinking water manufacturing company has been undertaken
and the information regarding different items produced, number of units produced, labour hour required, profit
earned for each of the items has been collected. Linearity of the information is checked with help of coefficient
of determination. Profit function is generated along with the underlying set of constraints and the convexity is also
verified and eventually a linear programming problem of the company is formulated. Solving the problem by the
powerful simplex technique in Excel-Solver, the maximum profit as well as the optimum number of items to be
produced are obtained.

Keywords: linear programming; optimization; simplex technique; convexity; coefficient of determination

1. Introduction:

The linear programming (LP) is a mathematical method to provide an optimal solution for the problems where
objective and requirements are both linear. The Linear Programming technique came to limelight during Second
World. Scientists were involved to find out a scientific way to win over the other countries of the world. The same
technique is now used in many industries for how many to produce in order to earn maximum profit. Most
common methods in this regard are graphical and simplex methods.
Linear Programming was first introduced by Leonid Kantorovich in 1939. He developed the earliest linear
programming problems that were used by the army during WWII in order to reduce the costs of the army and
increase the efficiency in the battlefield. The method was a secret because of its use in war-time strategies, until
1947 when George B. Dantzig published the simplex method and John von Neuman developed the theory of
duality. After WWII, many industries began adopting linear programming for its usefulness in planning
optimization.
Dantzig's original linear programming example was to find the best assignment of 70 people to 70 jobs. In
order to select the best assignment requires a lot of computing power; the number of possible configurations
exceeds the number of particles in the observable universe. However, by posing the problem as a linear program
and applying the simplex algorithm, it takes only a moment to find the optimum solution. The theory behind linear
programming drastically reduces the number of possible optimal solutions that must be checked.
Linear programming (LP) may be defined as the problem of maximizing or minimizing a linear function
that is subjected to linear constraints. The constraints may be equalities or inequalities. The Optimization problems
involve the calculation of profit and loss. Linear programming problems are an important class of optimization
problems that helps to find the feasible region and optimize the solution in order to have the highest or lowest
value of the function.
In other words, linear programming is considered as an optimization method to maximize or minimize the
objective function of the given mathematical model with the set of some requirements which are represented in
the linear relationship. The main aim of the linear programming problem is to find the optimal solution.
Linear programming is the method of considering different inequalities relevant to a situation and
calculating the best value that is required to be obtained in those conditions. Some of the assumptions taken while
working with linear programming are:
• The number of constraints should be expressed in the quantitative terms.
• The relationship between the constraints and the objective function should be linear.
• The linear function (i.e., objective function) is to be optimized.

JAFS|ISSN 2395-5554 (Print)|ISSN 2395-5562 (Online)|Vol 8(2)|December 2022 81

Journal of Applied and Fundamental Sciences

The basic components of the LP are as follows:

• Decision Variables, Constraints, Data, Objective Functions.
Our project is based on the following companies:
• Ozonised Purell Packaged Drinking Water Company

1.1 literature review:

To discuss the application of Linear Programming in various practical situations, a Literature review has been
done as follows:
Render et. al. in the year 2012 [1], puts an emphasis on model building and computer applications to show how
the techniques presented in the text are used in business. This text’s use of software also allows instructors to
focus on the managerial problem, while spending less time on the mathematical details of the algorithms. In the
eleventh edition, Excel 2010 has been incorporated throughout the text and an even greater emphasis on modelling
is provided.
Hasib and Hasan 2013 [2], in his research paper they represented formulating linear programming in real
life problem using computer techniques AMPL and LINGO.
Taha in the year 2011 [3] ,the first formal activities of OR were initiated in England during World War II,
when a team of British scientists set out to make scientifically based decisions regarding the best utilization of
war materiel. After the war, the ideas advanced in military, they have taken note of this important statement during
the preparation of the ninth edition, making every effort to introduce the art of modelling in OR.
Taylor in the year 2013 [4], a simple straightforward approach to modelling and solution techniques.
Introduction to Management Science shows how to approach decision-making problems in a straightforward,
logical way. Through the use of clear explanations and learn how to solve problems and make decisions based on
the results. The eleventh edition reflects latest version of Excel and provides many new problems for instructors
to assign.
Murty in the year 1983 [5], The goal of this journal is to provide a central forum for the distribution of
timely information about network problems, their design and mathematical analysis, as well as efficient algorithms
for carrying out optimization on networks. The nonstandard modelling of diverse processes using networks and
network concepts is also of interest. Consequently, the disciplines that are useful in studying networks are varied,
including applied mathematics, operations research, computer science, discrete mathematics and economics.
Dantzing and Thapa in the year 2003 [6], Linear programming represents one of the major applications of
mathematics to business, industry, and economics. It provides a methodology for optimizing an output given that
is a linear function of a number of inputs. George dantazing is widely regarded as the founder of the subject with
his invention of the simplex algorithm in the 1940's. This second volume is intended to add to the theory of the
items discussed in the first volume. It also includes additional advanced topics such as variants of the simplex
methods, interior point methods.
Vazirani in the year 2001 [7], This book covers the dominant theoretical approaches to the approximate
solution of hard combinatorial optimization and enumeration problems. It contains elegant combinatorial theory,
useful and interesting algorithms, and deep results about the intrinsic complexity of combinatorial problems. Its
clarity of exposition and excellent selection of exercises will make it accessible and appealing to all those with a
taste for mathematics and algorithms.
By George and B. Dantzig in the year 1997 [8], the Story about How It Began: Some legends, a little about
its historical sign-cancel, and comments about where its many mathematical programming extensions may be
headed. Industrial production, resources in the economy, the exertion of military or in a war—all require the
coordination of interrelated activities. What these complex undertakings share in common is the task of
constructing a statement of actions to be.
Neter and Nachtchaim in the year 1996 [9], This article presents necessary and sufficient conditions to be
satisfied by the best linear unbiased predictor of future observations in the general linear model in order to have a
simple form. Under these conditions, the predictors have an expression similar to that in the uncorrelated case and
some parameters related to the co variances between some observations need not to be known.
Kudragavtsen in the year 1981 [10], now a days mathematical method is widely applied in planning of
natural economy, organization of industry control, business decision, transportation, engineering,
telecommunications, elaboration of military operations etc. From the general point of view, the problems of
control and planning are usually reduced to a choice of a certain system of numerical parameters or a function
ensuring the most effective achievement of the pre planed aim (optimum plan) with the limited possible resources
taken into account. To estimate the effectiveness of a plan, introduce the plan quantity index expressed in term of
the plan characteristics and attaining the extremism value for an optimal plan. For the large number of practically
interesting problems the objective function is expressed linearly in term of plan characteristics, the permissible

JAFS|ISSN 2395-5554 (Print)|ISSN 2395-5562 (Online)|Vol 8(2)|December 2022 82

Journal of Applied and Fundamental Sciences

values of the parameters also obeying linear equalities or inequalities.

1.2 Objectives:

On the basis of the literature studied, the following objectives are carried out in this project--
• Optimizing the number of units per item manufactured by the companies.
• Optimizing the profit

2. Materials and Methods:

2.1 Materials:

This case study is based on the Ozonised Purell Packaged Drinking Water Company. This company was
established on 2016 and it is situated in Nalbari, Assam, India. Products prepared by this company are: 20-liter,
2-liter, 1-liter and Half-liter water bottles. Information gathered from this company are displayed in Table 2.1.

Table 2.1: Materials of Ozonised Purell Packaged Drinking Water Company:

PRODUCT 20 LITER BOTTLE 2 LITER BOTTLE

Size Profit Labour Metal Size Profit Labour Metal

YEAR hour Hour

2016 15000 100200 5000 307500 25000 10000 834 52500

2017 16500 124000 5500 338250 28000 136000 930 58800

2018 18000 150000 6000 369000 35000 180000 1166 73500

2019 21000 170000 7000 430500 45000 220000 1500 94500

2020 25000 220000 8334 512500 50000 250000 1666 105000

PRODUCT 1 LITER BOTTLE HALF LITER BOTTLE

2016 Size Profit Labour Metal Size Profit Labour Metal

Hour Hour
2017 30000 80000 500 36000 45000 70000 750 27000

2018 40000 120000 668 48000 54000 120000 900 32400

2019 60000 155000 1000 72000 60000 154000 1000 36000

2020 75000 230000 1250 90000 74000 190000 1240 44400

80000 250000 1340 96000 90000 230000 1500 54000

JAFS|ISSN 2395-5554 (Print)|ISSN 2395-5562 (Online)|Vol 8(2)|December 2022 83

Journal of Applied and Fundamental Sciences

2.2 Methodology:

• Define Decision variables.

• Formulation of objective function.
• Formulation of the constraints.
• Non-negative constraints.
The linear relationship should be examined and this is done with the help of R2 value (coefficient of
determination). This value should be reported for a relationship between quantity of production of each kind of
products and profits, and quantity of production and labour hours, quantity of production and metal, and time
progressing and size of production. Calculation and examination of R2 is done in MS-EXCEL which is followed
by F-ratio. Practically speaking, a higher R2 indicates that a linear model fits the data well; therefore, if the R2 is
high, a large F statistic should follow, indicating strong evidence that at least some of the coefficients are non-
zero. The p-value of the F-statistic is also implemented for double verification of the existence of linear
relationship. The examination is done at 5% level of significance. If the p-value is less than the level of
significance, then the model fits the linearity.
For obtaining the objective function, coefficients of the profit per unit for each of the products is taken from
regression analysis table. For the constraints, the coefficients of each of the products are considered. t-statistic and
p-value conform the normality of the data.
To check the convexity of a function, we need to examine the second-order derivative conditions.
Specifically, for a function 𝑓: 𝑅𝑛 → 𝑅 to be convex, its Hessian matrix must be positive semidefinite.
The Hessian matrix H for a function f(x) is given by the matrix of second partial derivatives:

𝜕2𝑓
𝐻𝑖𝑗 =
𝜕𝑥𝑖 𝜕𝑥𝑗

3. Results and discussion:

3.1 Results: mathematical formulation:

3.1.1 Status of Relationship:

Status of relationships between product and profit, product and labour hour, product and metal required and impact
of time of progressing and unit of the products are displayed in table 3.1.1.

JAFS|ISSN 2395-5554 (Print)|ISSN 2395-5562 (Online)|Vol 8(2)|December 2022 84

Journal of Applied and Fundamental Sciences

Table 3.1.1: Status of Relationship

RELATIONSHIP R2 F-value p-value Status

The impact of unit of produced 20 litter bottle 0.98221 165.6332 0.001012 Linear
and profit
The impact of unit of produced 2 litter bottle 0.809797 12.7726 0.037449 Linear
and profit
The impact of unit of produced 1 litter bottle 0.969734 96.12185 0.002255 Linear
and profit
The impact of unit of produced Half litter 0.95639 65.7912 0.003912 Linear
bottle and profit
The impact of unit of produced 20 litter bottle 1 1.9E+08 8.4E-13 Linear
and labour Hour
The impact of unit of produced 2 litter bottle 0.999983 173530.1 3.05E-08 Linear
and labour Hour
The impact of unit of produced 1 litter bottle 0.999954 64902.75 1.33E-07 Linear
and labour Hour
The impact of unit of produced Half litter 0.999907 32357.92 3.79E-07 Linear
bottle and labour Hour
The impact of unit of produced 20 litter bottle 1 6.96E+32 1.2E-49 Linear
and metal
The impact of unit of produced 2 litter bottle 1 3.94E+32 2.82E-49 Linear
and metal
The impact of unit of produced 1 litter bottle 1 2.13E+32 7.09E-49 Linear
and metal
The impact of unit of produced Half litter 1 7.26E+34 1.13E-52 Linear
bottle and metal
The impact of time of progressing and unit of 0.949763 56.71654 0.004853 Linear
20 litter bottle
The impact of time of progressing and unit of 2 0.97333 109.4878 0.001863 Linear
litter bottle
The impact of time of progressing and unit of 1 0.969415 95.08696 0.002291 Linear
litter bottle
The impact of time of progressing and unit of 0.967072 88.1068 0.002561 Linear
Half litter bottle

3.1.2. Status of Objective Function:

Status of objective function is shown in table 3.1.2.

Table 3.1.2: Status of objective function

Products Coefficient (Profit per unit) t –value p-value
20 Litter bottle 11.436 12.87 0.001
2 Litter bottle 7.859 3.574 0.037
1 Litter bottle 3.274 9.804 0.002
Half Litter bottle 3.417 8.111 0.004

Table 3.1.2 shows the profit per unit (coefficients) of 20 litter-bottle, 2 litter-bottle, 1 litter-bottle and half-litter-
bottle along with their significance at 5% level of significance. All the p-values are less than the level of
significance and hence they significant for consideration. Thus, the objective function thus obtained is:
𝑀𝑎𝑥. 𝑝𝑟𝑜𝑓𝑖𝑡 = 11.436𝑥1 + 7.859𝑥2 + 3.274𝑥3 + 3.417𝑥4 − − − −(3.1)

JAFS|ISSN 2395-5554 (Print)|ISSN 2395-5562 (Online)|Vol 8(2)|December 2022 85

Journal of Applied and Fundamental Sciences

3.1.3 Status of Constraints:

Status of Constraints is presented in table 3.1.3.

Table 3.1.3: Status of constraints

Products Coefficient t-value p-value
20 Litter bottle 0.333 Labour 13793.95 0.000
2 Litter bottle 0.033 Labour 416.5694 0.000
1 Litter bottle 0.017 Labour 2 54.7602 0.000
Half Litter bottle 0.017 Labour 179.8831 0.000
20 Litter bottle 20.5 Metal 2.64E+16 0.000
2 Litter bottle 2.1 Metal 1.98E+16 0.000
1 Litter bottle 1.2 Metal 1.46E+16 0.000
Half Litter bottle 0.6 Metal 2.69E+17 0.000

The constraints thus obtained here are:

𝑥1 ≤ 26450
𝑥2 ≤ 56700
𝑥3 ≤ 97500
𝑥4 ≤ 97600 − − − −(3.2)
0.333𝑥1 + 0.033𝑥2 + 0.017𝑥3 + 0.017𝑥4 ≤ 13966.8
20.5𝑥1 + 2.1𝑥2 + 1.2𝑥3 + 0.6𝑥4 ≤ 836855
𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 ≥ 0 }

3.1.4 Convexity:

In our case, the objective function is 𝑓(𝑥) = 11.436𝑥1 + 7.859𝑥2 + 3.274𝑥3 + 3.417𝑥4 . Since this is a linear
function, all the second partial derivatives will be zero, and the Hessian matrix will be a matrix of zeros. A matrix
of zeros is always positive semidefinite, so the function is convex.
In summary, the function 𝑓(𝑥) = 11.436𝑥1 + 7.859𝑥2 + 3.274𝑥3 + 3.417𝑥4 is convex.

3.1.5 Solution:

To know the result of objective function (3.1) with respect to the set of constraints (3.2), MS-Excel Solver has
been incorporated. Table 3.1.5 portrays below the complete result of the linear programming problem consists of
the function (3.1) and set of restrictions (3.2).

Table 3.1.5: Ozonised Purell Packaged Drinking Water Purification Results:

Variables Solution
𝑥1 26363.36
𝑥2 56700
𝑥3 97500
𝑥4 97600
Objective

Maximize 1399811
Constraints
Inequality RHS
1 26363.36 ≤ 26450
2 56700 ≤ 56700
3 97500 ≤ 97500
4 97600 ≤ 97600
5 13966.8 ≤ 13966.8
6 835078.9 ≤ 836855

JAFS|ISSN 2395-5554 (Print)|ISSN 2395-5562 (Online)|Vol 8(2)|December 2022 86

Journal of Applied and Fundamental Sciences

7 26363.36 ≥ 0
8 56700 ≥ 0
9 97500 ≥ 0
10 97600 ≥ 0

3.2 Discussion:

In section 3.1.1, the status of relationship and linearity is examined and is found linear. These are verified with
the application of 𝑅2 and p-value. In section 3.1.2, coefficients of the objective function are obtained and finally
the objective function is generated which is shown in function (3.1). in section 3.1.3, the coefficients of the
constraints of the problem are obtained and eventually the set of constraints is formulated which is shown in
function (3.2). Clubbing function (3.1) and (3.2), the formulation of the LP problem is framed. It is a simple LP
situation. It can be solved by various software. Excel Solver is one of them. Adopting simplex method in Excel
Solver in section 3.1.5, the problem is solved. The objective function gives an amount of Rs. 1399811 and 𝑥1 =
26363, 𝑥2 = 56700, 𝑥3 = 97500 and 𝑥4 = 97600. These results help us to infer that maximum profit for the
year 2021 has been found to be Rs. 1399811, units of 20 litre bottle to be produced is 26363, unit of 2 litre bottle
to be produced is 56700, unit of 1 litre bottle to be produced is 97500 and unit of Half litre to be produced is
97600.

4. Conclusion:

In this case study, a company has been taken for studying the application of linear programming problem to
maximize the profit and determine the number of units of each product to be produced. In section 1.2, the
objectives are framed. To meet the objectives, the information is processed in section 3.1.1, 3.1.2 and 3.1.3. in
section 3.1.4, convexity of the problem is justified. Section 3.1.5 brings out the solution of the entire problem.
This has been discussed in section 3.2. From the discussion, the following conclusions are inferred out:
• Information from the company are collected.
• Simplex method is adopted in Excel solver.
• Maximum profit for the year 2021 found to be: Rs. 1399811.
• Number of 20 litre bottle to be produced is 26363.
• Number of 2 litre bottle to be produced is 56700.
• Number of 1 litre bottle to be produced is 97500.
• Number of Half litre to be produced is 97600.

In this study, the company undertaken falls within a certain jurisdiction only. The same can be extended to work
for all small-scale as well as large scale industries and students of mathematics can take it as a source of income
by rendering the service to the industries.

References:

[1] S. Dey and M. Baruah, Journal of Applied and Fundamental Sciences 10, 100, 2015.
[2] M. Baruah, S. Dey and Kanan K. Ojah, A Text Book on Advances in Fundamental and Applied Sciences,
Assam Don Bosco University Press Inc., Guwahati, 2015.
[3] S. Boruah, Pure Sciences, In H. Bharali, B. Sinha and M. Biswas, (ed.), A Text Book on Advances in Physics
and Chemistry, Assam Don Bosco University Press Inc., Guwahati, 2015.
[4] H. Bharali, (ed.), A Text Book on Mathematics, Assam Don Bosco University Press Inc., Guwahati, 2015.
[5] B. Sinha and M. Biswas, A Conference Paper, In S. Dey, (ed.), Proceedings of National Conference NABSET
2015, Assam Don Bosco University Press Inc., Guwahati, 2015.
[6] N. Kapil, A Study of Fundamental and Applied Sciences, Ph.D. thesis, Assam Don Bosco University, 2015.

[1] B. Render, R. M Stair. and M. E. Hanna, Quantitative Analysis for Management, Newyork, 2012.
[2] Md. Hasib Uddin Mollah and M. Babul Hasan, Art of Formulation and LPs and IPs m Real life Problem,
Dhaka3, 2013.
[3] H. Taha, Operation Research: an introduction, New york, 2011.
[4] B. Taylor, Introduction to management science, New York: Pearson, 2013.
[5] K. G. Murty, Linear Programming, New York: John Wiley and Sons.Inc, 1983.
[6] G. B Dantzing & M. N Thapa, Linear Programming2, Theory and Extension. Springer-Verlag, 2003.

JAFS|ISSN 2395-5554 (Print)|ISSN 2395-5562 (Online)|Vol 8(2)|December 2022 87

Journal of Applied and Fundamental Sciences

[7] V. Vazirani, Approximation Algorithms, Spring-Verlag, 2001.

[8] G. B. Dantzing, & M. N Thapa, Linear Programming 1 Introduction, Springer-Verlag., 1997.
[9] K. Neter, & W. Nachtchaim, Applied Linear Statistical Method, New York: McGraw-Hill, 1996.
[10] V. A. Kudragavtsen, & B. P. Demaidovich, A Brief Course of Higher Mathematics, Moscow: Mir Publishers,
1981.

JAFS|ISSN 2395-5554 (Print)|ISSN 2395-5562 (Online)|Vol 8(2)|December 2022 88