Special Issue - (AAM - 2019) Advances in Applied Mathematics (AAM - 2019)

ISSN No:-2456-2165

Study on Simplex Method and Design &

Development of Application for Solving Simplex
Problem using Python
R Vrishab Jain Vishal Patil
Department of Computer Science and Engineering Department of Mathematics
SET – Jain Deemed-to-be University SET – Jain Deemed-to-be University
Bangalore, India Bangalore, India

Radha Gupta Ravinder Singh Kuntal

Department of Mathematics Department of Mathematics
DayanandaSagar College of Engineering Jain University
Bangalore, Karnataka, India Bangalore, Karnataka, India

Abstract:- The aim of this study is to develop application 𝑀𝑎𝑥 𝑍 = 𝐶𝑋

using python language which solves the simplex problem
and this will help the user to understand how each step Subject to 𝐴𝑋 = 𝑏
follows in simplex problem until the optimum solution is And 𝑋≥0
reached. Here we can optimize the objective function for
minimization as well as maximization and linear Where C is the cost matrix, X is variable matrix, A is
constraints with n number of variables. Application coefficient matrix, is the right hand side of the constraints
developed is user friendly; python code takes input from [3]. Real world problems like production, sales and
the user in the console. Every input has a proper labelled management, are solved using simplex algorithm. This study
question and the user has to enter the values of what is works in two phases’ first one, finding appropriate method
asked, and it shows the solution in excel with all iteration of solving a LPP, which has different types of constraints(≤
including key column and key row. The simplex , ≥, =). In the Second phase converting this method into a
algorithm using python can extend to solve real world code using python so that method can be used for all inputs.
problems like least cost formulation and maximization of
profit. Python is a high level programming language used
almost everywhere. It was created by Guido van Rossum at
Keywords:- Simplex Problem, Python Application, Centrum Wiskunde & Informatica (CWI) in Netherlands and
Optimization, Excel. was first released in 1991. It supports multi-programming
paradigms, which include object oriented, procedural and
I. INTRODUCTION functional and many more. Python has developed a lot
throughout the year since its first release, from being a
The introduction of paper contains the nature of ‘hobby’ project to Guido van Rossum to the one of the most
research work, purpose of work, and the contribution of this useful languages. Python is used in Data Science, Machine
paper. Learning, Web servers and many more places. Python is one
of the easiest language to learn and it being an open source
Simplex method is one of the approaches to solve software anyone in the python community and contribute to
linear programming problem. This method invented by its development like ‘xlsxwriter’, ‘flask’ etc. making it
George Dantzig in 1947. Simplex is very efficient method highly extensible. Today anyone can create useful packages
[1]. for python so that the other programmers do not have to do
the same thing twice but just use it how they want. Python is
Simplex in one dimension is a line segment connecting meant to be an easily readable language and it is possible
two points, in two dimensions it is a triangle formed by the only due to its formatting. It contains English keywords like
points. A three dimensions simplex is the four-sided ‘in’, ‘as’ etc. and the indentation used by python is
pyramid having four corners. Simplex finds the best corner whitespace indentation where increase or decrease in
of the feasible region. It usually starts with the corner point indentation is used to determine the opening or closing block
which has no values then it moves to neighborhood point rather the usually used curly braces in many of the
which improves the solution simplex repeats the process languages. Python can also act as a cross compiler to other
until gets the optimum solution that is no further more languages like Java, R, JavaScript, C/C++, VHDL in the
improvements [2]. Standard form of simplex method have form of JPython, RPython, Pyjs, Cython and MyHDL
three main requirements one is objective function, should be respectively. Python’s design and concept has influenced
maximization, second all constraints to be linear and, third many other languages which include Cobra, Go, Julia,
all variables to be non-negative. The general form is as Kotlin, Ruby, Swift and many more.
shown below,

AAM19MY09 www.ijisrt.com 39
Special Issue - (AAM - 2019) Advances in Applied Mathematics (AAM - 2019)
ISSN No:-2456-2165
II. MATERIALS & METHODS 2. If the constraint inequalities are ′ ≥’ then subtract
surplus (𝑆) variable and add artificial variable(𝐴) to the
The simplex method can be easily applies to real world left hand side of the inequality and write it as equation.
problems. The problem is converted into mathematically 3. If the constraint inequalities is′ =’ then add artificial
linear model to keep it into standard form. Simplex method variables (𝐴) to the left hand side of the inequality and
procedure which is been followed in this study is given write it as equation.
below with an example 4. If in case right hand side is negative then multiply the −1
throughout and make it positive and then follow rules(1-
Example: 𝑀𝐼𝑁 𝑍 = 𝑥1 + 𝑥2 + 3𝑥3 3)
5. In the objective function coefficient of slack variable,
Subject to constraints. surplus variable is 0 and artificial variable is -1
3𝑥1 + 2𝑥2 + 𝑥3 ≤ 3
2𝑥1 + 𝑥2 + 2𝑥3 ≥ 3 𝑀𝑎𝑥 𝑍 = −𝑥1 − 𝑥2 − 3𝑥3 + 0𝑆1 + 0𝑆2 − 𝐴1 − 𝐴2
𝑥1 + 𝑥2 + 𝑥3 = 3/2 3𝑥1 + 2𝑥2 + 𝑥3 + 𝑆1 = 3
𝑥1 , 𝑥2 , 𝑥3 ≥ 0 2𝑥1 + 𝑥2 + 2𝑥3 − 𝑆2 + 𝐴1 = 3
𝑥1 + 𝑥2 + 𝑥3 + 𝐴2 = 3/2
Step 1: Make sure that objective function is Maximization.
If it is minimization then convert it into maximization by Step 3: Non-negativity constraints are
multiplying -1 and write 𝑀𝑎𝑥 𝑍 = 𝑀𝑖𝑛(−𝑧). 𝑥1 , 𝑥2 , 𝑥3 , 𝑆1, 𝑆2, , 𝐴1, 𝐴2, ≥ 0

𝑀𝑎𝑥 𝑍 = −𝑥1 − 𝑥2 − 3𝑥3 Step 4: Initial basic feasible solution can be found by
equating 𝑛 − 𝑚 variable to zero (𝑛 = number of variable
Step 2: Convert the inequalities to equalities by employing and 𝑚 = number of constraints). That is 6 − 3 = 3 variables
following rule. are equated to zero those are 𝑥1 = 𝑥2 = 𝑥3 = 0, therefore
we Basic variables 𝑆1 = 3, 𝐴1 = 3, 𝐴2 = 3/2. Write the
1. If the constraint inequalities is′ ≤’ then add slack linear problem in tabular as shown below,
variables (𝑆) to the left hand side of the inequality and
write it as equation.

𝑪𝒋 -1 -1 -3 0 0 -1 -1

𝒂𝒊 B.V 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝑺𝟏 𝑺𝟐 𝑨𝟏 𝑨𝟐 R.H.S Ratio

𝟑
0 𝑺𝟏 3 2 1 1 0 0 0 3 =𝟏
3 𝟑

𝟑
-1 𝑨𝟏 2 1 2 0 -1 1 0 3 = 𝟏. 𝟓
𝟐
𝟑
-1 𝑨𝟐 1 1 1 0 0 0 1 3/2 = 𝟏. 𝟓
𝟐
𝒁𝒋 -3 -2 -3 0 1 0 0 −𝟗/𝟐
Table 1:- Table for example input

In the table, 𝐶𝑗 are coefficients of objective function pivot column ratios is found, which is shown in the right
and 𝑎𝑖 are coefficients of basic variable in the objective hand side. The row, which has, has lowest ratio that will be
function. 𝑍𝑗 − 𝐶𝑗 is calculated by the formula ∑𝟑𝒊=𝟏 𝒂𝒊 𝒙𝒊𝒋 𝑗 = pivot row, and the variable is departing variable (𝑆1 ). and
1, 2 … .7. coinciding number is called pivot element(3).

For example,,[(𝟎𝑿𝟑) + (−𝟏𝑿𝟐) + (−𝟏𝑿𝟏)] = −𝟑.

Using this formula all the values are calculated and if all the
values are positive then optimality is reached. For artificial Step 5: The next iterative table all the pivot row element are
variable, it is kept zero. If the LPP contain all constraints’ divided by the pivot element and remaining rows can be
less than equal to zero then ∑3𝑖=1 𝑎𝑖 𝑥𝑖𝑗 − 𝐶𝑗 will be found. If found by using the formula
a value is less than zero, it means that variable has not
𝑛𝑒𝑤 𝑟𝑜𝑤 =
reached its optimal value, as we seen in the above table there
are two negative values, select the most negative value, the
𝑜𝑙𝑑 𝑟𝑜𝑤 − (𝑃𝑟𝑜𝑑𝑢𝑐𝑡 𝑜𝑓 𝑒𝑙𝑒𝑚𝑛𝑡 𝑜𝑓 𝑝𝑖𝑣𝑜𝑡 𝑐𝑜𝑙𝑢𝑚𝑛 & 𝑝𝑖𝑣𝑜𝑡 𝑟𝑜𝑤)
column which consists that value called as pivot column and
the variable is entering variable(𝑥1 ). For each value in the 𝑝𝑖𝑣𝑜𝑡 𝑒𝑙𝑒𝑚𝑒𝑛𝑡

AAM19MY09 www.ijisrt.com 40
Special Issue - (AAM - 2019) Advances in Applied Mathematics (AAM - 2019)
ISSN No:-2456-2165
The above picture depicts how the excel sheet looks
The process continues until the optimality condition after doing the operations on the entered equations. The red
reached that is all 𝒁𝒋 − 𝑪𝒋 ≥ 𝟎. cells contain the value, which is least in the Zj-Cj row, and
the yellow cells contain the key values and the green cells
III. SIMPLEX APPLICATION USING PYTHON contain the values of the variables and Z after doing the
simplex method.
The application is created using the programming
language python. The python code takes input from the user B. How does it work??
in the console. Every input has a proper-labelled question The first thing the code does is, it takes the number of
and the user has to enter the values of what is asked. The equations and variables as input, so now we know how
input console will look like the figure shown below (The many equations are there and the number of variables and
given input is for the above specified example). slack variables. After this, we find out the least value, which
is there in the Zj-Cj row. That is done by the following code:

Fig 3:- Code to find key column

Now we know which value is least in the Zj-Cj row i.e.

we know the key column and now we have to find the key
row, which is done the following code:

Fig 1:- Input example

In the above picture depicts how the program takes the Fig 4:- Code to find key row
inputs. The program first asks the user the number of
equations and number of variables (i.e. x1, x2, x3, etc.). From the above code, we have found the least row and
Then the program asks the user to enter the coefficients of the key value. The next step is to generate the next iteration.
each variable in each question. Then the user enters the After generating the next iteration the above steps for the
R.H.S value of the equation i.e. the “xb” value in the above simplex method is followed and when the answer is found
picture. Then the user is asked if the equation inequality the, it is shown in the excel file.
which are l(<=) less than or equal to inequality (or) g(>=)
greater than or equal to inequality (or) e(=) equality and the C. Examples:
input is case insensitive. After taking all the equation, the
user enters the coefficient of variable in the objective 1)
function. After entering these values and the code generates 𝑀𝑎𝑥 𝑍 = 2𝑥1 + 4𝑥2 + 3𝑥3
the excel file which is stored in the documents folder in the
user system. Subject to constraints,

The excel file contains all the iteration of the above 3𝑥1 + 4𝑥2 + 2𝑥3 ≤ 60
entered equation using simplex method. 2𝑥1 + 𝑥2 + 2𝑥3 ≤ 40
𝑥1 + 3𝑥2 + 2𝑥3 ≤ 80
A. Output 𝑥1 , 𝑥2 , 𝑥3 ≥ 0

Fig 2:- Excel file Fig 5:- Excel output for above example

AAM19MY09 www.ijisrt.com 41
Special Issue - (AAM - 2019) Advances in Applied Mathematics (AAM - 2019)
ISSN No:-2456-2165
2) FUTURE SCOPE
𝑀𝑖𝑛 𝑍 = 12𝑥1 + 20𝑥2
This application can be used for multiple purposes in
Subject to constraints future. It can use for feed formulation, linear programming
and integer programming. The application currently runs on
6𝑥1 + 8𝑥2 ≥ 100 a console, but in future, it can be upgraded with a
7𝑥1 + 12𝑥2 ≥ 120 completely user-friendly interface to make finding solution
𝑥1 , 𝑥2 ≥ 0 easier.

REFERENCES

[1]. https://www.imse.iastate.edu/files/2015/08/Explanation-
of-Simplex-Method.docx
[2]. J. Reeb., S. Leavengood., operations research, October
1998. Performance excellence in wood products
industry
[3]. S. Kalavathy., 2013. Operations research 4th Edition.
Vikas publishing house pvt.ltd.
[4]. https://maklaboratory.com/

Fig 6

IV. RESULT & DISCUSION

There are many applications available for finding the

results for simplex. In this section, we talk about how the
developed application is different and simple to analyze the
steps of the procedure. In the existing software, there are not
many options such as many of them do not compute
minimization of the objective function. To overcome this
challenge, we have developed this application we can solve
all kinds of linear programming problems.

 System Requirements:
 Any Windows system with Microsoft Excel can be used.
No particular
 Requirement for hardware or RAM.

V. CONCULSION

In this paper we have developed a completely

functional windows application which works for n no. of
equations and variables and shows the output in an excel
sheet on the same system. The application can work for any
type of equation using less than or equal to or greater than or
equal to or equal to equations and it finds the optimal
solution using simplex method matrix form. The excel sheet
contains each iteration formed during the course of finding
the solution and the key column value and key row value are
also show in a red and green boxes respectively. The key
column coefficient and the key row coefficient are also
displayed in the excel for the better understanding for the
user and the solution i.e. the optimal solution for the given
variables are displayed after all the iteration in green boxes.

AAM19MY09 www.ijisrt.com 42