International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP.

16-25, 2024

A Study of Systems of Neutrosophic Linear Equations

Maissam Jdid 1,*, Florentin Smarandache 2
1
Faculty of Science, Damascus University, Damascus, Syria
2
University of New Mexico ،Mathematics, Physics and Natural Sciences Division
705 Gurley Ave., Gallup, NM 87301, USA
Emails: [email protected]; [email protected]

Abstract

Operations research methods are among the modern scientific methods that have occupied a prominent place
among the mathematical methods used in planning and managing various economic and military activities. They
have been able to help specialists in developing ideal plans in terms of costs, production, storage, or investment of
human energies. One of its most important methods is the method Linear programming, which was built based on
the sets of linear equations that represent the constraints for any linear model. Based on the methods for solving
the systems of linear equations, researchers were able to prepare algorithms for solving linear models, such as the
direct Simplex algorithm and its modifications. After the emergence of neutrosophic science, we found that
research methods had to be reformulated. Operations using the concepts of this science, and as a basis and
foundation for neutrosophic linear programming. In this research, we will reformulate the systems of linear
equations and some methods for solving them using the concepts of neutrosophic to be a basis for any study
presented in the field of neutrosophic linear programming.

Keywords: Operations research; linear programming; systems of linear equations; neutrosophic science; systems
of neutrosophic linear equations; methods for solving neutrosophic linear equations.

1. Introduction:
Through our endeavor to provide everything new and keep pace with scientific development and given the great
importance of the linear programming method as one of the methods of operations research, we found it necessary
to reformulate the systems of linear equations and some methods for solving them using the concepts of
neutrosophic, the science in which research and studies that use its concepts topped the research list in most
publishing houses. The reason for this is the accuracy of the results obtained in studies that use the concepts of this
science, some research [1-13].

Operations research methods are among the modern scientific methods that have occupied a prominent place
among the mathematical methods used in planning and managing various economic and military activities. They
have been able to help specialists in developing ideal plans in terms of costs, production, storage, or investment of
human energies. One of its most important methods is the method Linear programming, which relies on
mathematical models to express any practical system from real life or a proposed idea. Mathematical models
consist of an objective function and constraints. Classical operations research methods provided many algorithms
that help to find the optimal solution for these models. The construction of these algorithms was based on sentences
Linear equations. Based on the methods used to solve these sentences, such as the Jardin-Diver method and the
Simplex method, the direct Simplex algorithm and its modifications were developed to solve linear models, and
in view of the importance of the linear programming method and to keep pace with everything new in the field of
scientific research and after the emergence of neutrosophic science, the science that caused a great revolution. In
all fields of science, we found that it is necessary to reformulate the methods of operations research using the
concepts of this science, the science that takes into account all the changes that may occur to the issue under study
through the indeterminacy of the issue’s data. Therefore, we present in this research a study of sentences of linear
equations and some methods for solving them using the concepts of neutrosophic science to be a basis for any
study presented in the field of neutrosophic linear programming .we will take the matrix of constants on the right
side of these equations as neutrosophic values. Of the form 𝑁𝑏𝑗 = 𝑏𝑗 + 𝛿𝑗 where 𝛿𝑗 is indeterminacy and can take
one of the forms𝛿𝑗 ∈ {𝜇𝑖1 , 𝜇𝑖2 } or 𝛿𝑗 ∈ [𝜇𝑖1 , 𝜇𝑖2 ] , , also 𝑎𝑖𝑗 on the right side we will take it Neutrosophic values
16
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

i.e., 𝑁𝑎𝑖𝑗 = 𝑎𝑖𝑗 + 𝛾𝑖𝑗 where 𝛾𝑖𝑗 is indeterminacy and can take In one of the forms 𝛾𝑖𝑗 ∈ [𝜑𝑖𝑗1 , 𝜑𝑖𝑗2 ], or 𝛾𝑖𝑗 ∈
{𝜑𝑖𝑗1 , 𝜑𝑖𝑗2 }, which helps us obtain more accurate solution results that take into account all conditions.

2. Discussion:
Based on the study mentioned in the classical references on the systems of linear equations and methods for
solving them in some references [14-17].
We present the following study according to the concepts of neutrosophic science:
The systems of linear equations in which the number of equations equals 𝑚 and the number of variables in them
equals 𝑛 are given according to classical logic in the following general form:

𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 𝑏1

𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥𝑛 = 𝑏2
……………………………………
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚
It is written in the following matrix form:
𝐴. 𝑋 = 𝐵
Where:
a11 a12 … a1n b1 x1
a 21 a 22 … a 2n b x 2
A = [ ………………… ] B = [ 2] X = [..]
..
a m1 a m2 … a mn bm xn

Where 𝑎ij and 𝑏𝑖 are real numbers for all values of 𝑖 = 1,2, … , 𝑚 and 𝑗 = 1,2, … , 𝑛
We distinguished three cases of the systems of linear equations:
The first case: The number of equations is equal to the number of variables, i.e., 𝑚 = 𝑛
The second case: The number of equations is greater than the number of variables, i.e., 𝑚 > 𝑛
Third case: The number of equations is less than the number of variables, i.e., 𝑚 < 𝑛
Below we will present the systems of linear equations using the concepts of neutrosophic science, where we will
take the real numbers 𝑎ij and 𝑏𝑖 as neutrosophic numbers, that is, of the form 𝑁𝑏𝑖 and 𝑁𝑎𝑖𝑗 , indefinite values.
Perfectly determined, they can be any neighborhood of the real numbers 𝑎𝑖𝑗 and 𝑏𝑖 , written in one of the forms.
next:
𝑁𝑎𝑖𝑗 = 𝑎𝑖𝑗 + 𝜀𝑖𝑗 and 𝑁𝑏𝑖 = 𝑏𝑖 + 𝜇𝑖 where 𝜀𝑖𝑗 ∈ [𝜆1𝑖𝑗 , 𝜆2𝑖𝑗 ] or 𝜀𝑖𝑗 ∈ {𝜆1𝑖𝑗 , 𝜆2𝑖𝑗 } or otherwise, then the systems of
neutrosophic linear equations is written in the following form:
The systems of neutrosophic linear equations in which the number of equations equals 𝑚 and the number of
variables in them equals 𝑛are given in the following general form:
𝑁𝑎11 𝑥1 + 𝑁𝑎12 𝑥2 + ⋯ + 𝑁𝑎1n 𝑥𝑛 = 𝑁𝑏1
𝑁𝑎21 𝑥1 + 𝑁𝑎22 𝑥2 + ⋯ + 𝑁𝑎2n 𝑥𝑛 = 𝑁𝑏2
… … … … … … … … … … … … … … … ….
𝑁𝑎m𝑛 𝑥1 + 𝑁𝑎m𝑛 𝑥2 + ⋯ + 𝑁𝑎𝑚𝑛 𝑥𝑚 = 𝑁𝑏𝑚

In the following matrix form:

𝑁𝐴. X = 𝑁𝐵
Where:
𝑁𝑎11 𝑁𝑎12 … 𝑁𝑎1𝑛 𝑁𝑏1 x1
𝑁𝑎21 𝑁𝑎22 … 𝑁𝑎2𝑛 𝑁𝑏 x 2
𝑁𝐴 = [ ] 𝑁𝐵 = [ 2 ] X = [..]
………………… ..
𝑁𝑎𝑚1 𝑁𝑎𝑚2 … 𝑁𝑎𝑚𝑛 𝑁𝑏𝑚 xn
To find the general solution to the previous systems of equations, we examine them according to the three
aforementioned cases:
The first case: The number of equations is equal to the number of variables, i.e., 𝑚 = 𝑛.
We write the systems of equations as follows:
𝑁𝑎11 𝑥1 + 𝑁𝑎12 𝑥2 + ⋯ + 𝑁𝑎1n 𝑥𝑛 = 𝑁𝑏1
𝑁𝑎21 𝑥1 + 𝑁𝑎22 𝑥2 + ⋯ + 𝑁𝑎2n 𝑥𝑛 = 𝑁𝑏2
………………………………………………
17
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

𝑁𝑎𝑛1 𝑥1 + 𝑁𝑎𝑛2 𝑥2 + ⋯ + 𝑁𝑎𝑛𝑛 𝑥𝑛 = 𝑁𝑏𝑛

In matrix form:
𝑁𝐴 . 𝑋 = 𝑁𝐵
Where:
𝑁𝑎11 𝑁𝑎12 … 𝑁𝑎1𝑛 𝑁𝑏1 𝑥1
𝑁𝑎21 𝑁𝑎22 … 𝑁𝑎2𝑛 𝑁𝑏2 𝑥2
𝑁𝐴 = [ ] 𝑁𝐵 = [ ] 𝑋 = [ .. ]
………………… ..
𝑁𝑎𝑛1 𝑁𝑎𝑛2 … 𝑁𝑎𝑛𝑛 𝑁𝑏𝑛 𝑥𝑛
The proverbs matrix is a square matrix whose determinant is ∆𝑁 = |𝑁𝐴|
Here we distinguish two cases:
1- ∆𝑁 = 0 . This case results in two cases:
a. If ∆𝑁 = 0 and ∆𝑁 𝑥 ≠ 0 where ∆𝑁 𝑥 is the determinant resulting from the determinant of the matrix of
𝑗 𝑗
proverbs ∆𝑁 after replacing the column containing the proverbs of the unknown 𝑥𝑗 with the column of
constants (the values on the side The second of the equations) then the systems have no solution.
b. If ∆𝑁 = 0 and ∆𝑁 𝑥 = 0, this means that the systems of equations are not linearly independent, meaning
𝑗
that some of them are linearly related to each other. To address this case, we delete one of the two linearly
related equations, thus the number of equations decreases and becomes 𝑚′ where 𝑚′ = 𝑚 − 1 and 𝑚′ <
𝑛, which is identical to the second case that will be dealt with later.
c. If ∆𝑁 ≠ 0, that is, the systems of equations are linearly independent and the systems have a single solution,
which can be found in several ways. In this research, we study Gaussian- Jordan method, which is the
basis for the direct simplex algorithm that we use to obtain the optimal solution for linear models.
Gaussian- Jordan method for solving systems of neutrosophic linear equations in which 𝒎 = 𝒏:
To clarify the mathematical basis of this method, we write the equations in the following matrix form:
𝑁𝑎11 𝑁𝑎12 … 𝑁𝑎1𝑛 𝑥1 𝑁𝑏1
𝑁𝑎21 𝑁𝑎22 … 𝑁𝑎2𝑛 𝑥 2 𝑁𝑏
[ ] . [ . . ] = [ 2 ] (1)
………………… ..
𝑁𝑎𝑛1 𝑁𝑎𝑛2 … 𝑁𝑎𝑛𝑛 𝑥𝑛 𝑁𝑏𝑛
Or in the following abbreviated form:
𝑁𝐴 . 𝑋 = 𝑁𝐵 (2)
Since ∆𝑁 = |𝑁𝐴| ≠ 0, this means that the matrix 𝑁𝐴 has an inverse that is 𝑁𝐴−1 . We multiply both
sides of equation (2) by 𝑁𝐴−1 and we find:
𝑁𝐴−1 . (𝑁𝐴 . 𝑋) = 𝑁𝐴−1 . 𝑁𝐵
Hence, we get:
𝐼. X = 𝑁𝐵′
Which is written in the following detailed form:
1 0 0…0 𝑥1 𝑁𝑏1′
0 1 0…0 𝑥 2 𝑁𝑏 ′
[ ] . [ . . ] = [ 2 ] (3)
…………… ..
0 0 0…1 𝑥𝑛 𝑁𝑏𝑛′
This process is the basis of the Gaussian- Jordan method for solving a system of linear equations. In
order to convert Figure (1) to Figure (2), we follow the following steps:
1- We express Figure (1) in the following table:
Table No. (1): Table of equations
Variables 𝑥1 𝑥2 …. 𝑥𝑛 𝑁𝐵
Equations
1 𝑁𝑎11 𝑁𝑎12 …. 𝑁𝑎1𝑛 𝑁𝑏1
2 𝑁𝑎21 𝑁𝑎22 … 𝑁𝑎2𝑛 𝑁𝑏2
…. … … … … …
𝑛 𝑁𝑎𝑛1 𝑁𝑎𝑛2 … 𝑁𝑎𝑛𝑛 𝑁𝑏𝑛

2- We convert the matrix 𝑁𝐴 to the unit matrix 𝐼 by processing the lines of the table so that we make
all non-diagonal elements in all its rows equal to zero and the diagonal elements equal to one. When
we want to remove the variable 𝑥𝑠 from the equation 𝑡, we follow these steps:
a- We divide all the elements of row 𝑡 in which we want to make 𝑥𝑠 equal to one by 𝑁𝑎𝑡𝑠 , so 𝑥𝑠 becomes
equal to one and the other expressions change.
b- We set all elements of the column with 𝑥𝑠 (except row 𝑡) equal to zero.
c- We calculate the rest of the elements of the new table from the following two relationships:

18
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

′
𝑁𝑎𝑡𝑗 𝑁𝑎𝑖𝑗 𝑁𝑎𝑡𝑠 − 𝑁𝑎𝑖𝑠 𝑁𝑎𝑡𝑗
𝑁𝑎𝑖𝑗 = (𝑁𝑎𝑖𝑗 − 𝑁𝑎𝑖𝑠 )=
𝑁𝑎𝑡𝑠 𝑁𝑎𝑡𝑠
(4)
𝑁𝑏 𝑡 𝑁𝑏 𝑖 𝑁𝑎𝑡𝑠 − 𝑁𝑎𝑖𝑠 𝑁𝑏𝑡
𝑁𝑏𝑖′ = (𝑁𝑏𝑖 − 𝑁𝑎𝑖𝑠 )=
𝑁𝑎𝑡𝑠 𝑁𝑎𝑡𝑠 ]
The element 𝑁𝑎𝑡𝑠 is called the pivot element
From the previous processing we get the following table:
Table No. (2) Final solution table
Variables 𝑁𝑥1 𝑁𝑥2 …. 𝑁𝑥𝑛 𝑁𝐵′
Equations
1 1 0 …. 0 𝑁𝑏1′
2 0 1 … 0 𝑁𝑏2′
…. … … … … …
𝑛 0 0 … 1 𝑁𝑏𝑛′

Through the table, the systems of linear equations are written in the following matrix form:
𝐼. 𝑁𝑋 = 𝑁𝐵′
1 0 0…0 𝑥1 𝑁𝑏1′
0 1 0…0 𝑥 2 𝑁𝑏 ′
[ ] . [ . . ] = [ 2] ⟹
…………… ..
0 0 0…1 𝑥𝑛 𝑁𝑏𝑛′
𝑁𝑥1 𝑁𝑏1′
𝑁𝑥 𝑁𝑏 ′
[ 2] = [ 2]
.. ..
𝑁𝑥𝑛 𝑁𝑏𝑛′
⟹ 𝑁𝑥1 = 𝑁𝑏1′ , 𝑁𝑥2 = 𝑁𝑏2′ , … , 𝑁𝑥𝑛 = 𝑁𝑏𝑛′
The second case:
The number of equations is greater than the number of variables, i.e., 𝑚 > 𝑛:
In this case, we form a new system from the set of equations in which the number of equations is equal to the
number of variables by excluding a number of equations of m − n. Then we solve the new systems as we did in
the first case and replace the resulting solution in the equations that were excluded to ensure that they are satisfied.
The third case: The number of variables is greater than the number of equations, i.e., 𝑚 < 𝑛:
In this case, we are faced with a set of equations of the following form :
𝑁𝑎11 𝑥1 + 𝑁𝑎12 𝑥2 + ⋯ + 𝑁𝑎1m 𝑥𝑚 + ⋯ + 𝑁𝑎1n 𝑥𝑛 = 𝑁𝑏1
𝑁𝑎21 𝑥1 + 𝑁𝑎22 𝑥2 + ⋯ + 𝑁𝑎2m 𝑥𝑚 + ⋯ + 𝑁𝑎2n 𝑥𝑛 = 𝑁𝑏2
⟧ (5)
…………………………………………………………………
𝑁𝑎𝑚1 𝑥1 + 𝑁𝑎𝑚2 𝑥2 + ⋯ + 𝑁𝑎mm 𝑥𝑚 + ⋯ + 𝑁𝑎𝑚𝑛 𝑥𝑛 = 𝑁𝑏𝑚
Which is written in the following matrix form:
𝑥1
𝑥2
𝑁𝑎11 𝑁𝑎12 … 𝑁𝑎1𝑚 … … 𝑁𝑎1𝑛 𝑁𝑏1
..
𝑁𝑎21 𝑁𝑎22 … 𝑁𝑎2𝑚 … … 𝑁𝑎1𝑛 𝑁𝑏2 𝑥
𝑁𝐴 = [ ] 𝑁𝐵 = [ ] 𝑋= 𝑚
……………………………………… .. ..
𝑁𝑎𝑚1 𝑁𝑎𝑚2 … 𝑁𝑎𝑚𝑚 … … 𝑁𝑎𝑚𝑛 𝑁𝑏𝑛 ..
[ 𝑥𝑛 ]
In the following brief form:
𝑁𝐴(𝑚.𝑛) . 𝑋(𝑛.1) = 𝑁𝐵(𝑚.1) (6)
1- We partition the matrix 𝑁𝐴(𝑚.𝑛) into two matrices:
a- A square matrix of rank (𝑚. 𝑚) and we denote it as 𝑁𝐶(𝑚.𝑚) .
b- And a rectangular matrix of rank (𝑚. 𝑛 − 𝑚) and we denote it 𝑁𝐷(𝑚.𝑛−𝑚)
2- We partition the column matrix 𝑋(𝑛.1) into two matrices 𝑋 ′ (𝑚.1) and 𝑋 " (𝑛−𝑚.1) .
Then the systems of equations (5) are written in the following matrix form:

𝑋 ′ (𝑚.1)
[𝑁𝐶(𝑚.𝑚) , 𝑁𝐷(𝑚.𝑛−𝑚) ] . [ ] = 𝑁𝐵(𝑚.1) (7)
𝑋 " (𝑛−𝑚.1)
𝑁𝐶(𝑚.𝑚) . 𝑋 ′ (𝑚.1) + 𝑁𝐷(𝑚.𝑛−𝑚) . 𝑋 " (𝑛−𝑚.1) = 𝑁𝐵(𝑚.1)
From them we find that:
𝑁𝐶(𝑚.𝑚) . 𝑋 ′ (𝑚.1) = 𝑁𝐵(𝑚.1) − 𝑁𝐷(𝑚.𝑛−𝑚) . 𝑋 " (𝑛−𝑚.1) (8)
Assuming that |𝑁𝐶| ≠ 0, we multiply both sides in relation (8) by 𝑁𝐶 −1 and we find:
19
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
 International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

𝑁𝐶 −1 . 𝑁𝐶. 𝑋 ′ = 𝑁𝐶 −1 . (𝑁𝐵 − 𝑁𝐷. 𝑋 " )

𝐼. 𝑋 ′ = 𝑁𝐶 −1 . 𝑁𝐵 − 𝑁𝐶 −1 . 𝑁𝐷. 𝑋 " ) (9)
′
Assuming that 𝑁𝐶 . 𝑁𝐵 = 𝑁𝐵 and 𝑁𝐶 −1 . 𝑁𝐷 = 𝑁𝐷(𝑚.𝑛−𝑚)
−1 ′
we find that:
′ ′ ′
1 0 0…0 𝑥1 𝑁𝑏 1
′ 𝑁𝑑 11 𝑁𝑑 12 … 𝑁𝑑1(𝑛−𝑚) 𝑥𝑚+1
0 1 0…0 𝑥 2 𝑁𝑏 ′
𝑁𝑑21 ′ ′ ′
𝑁𝑑22 … 𝑁𝑑1(𝑛−𝑚) 𝑥𝑚+2
[ ].[ .. ] = [ 2]− . [ … ] (10)
…………… .. …………………………
0 0 0…1 𝑥𝑚 𝑁𝑏𝑚 ′ ′ ′
[𝑁𝑑𝑚1 𝑁𝑑𝑚2 … 𝑁𝑑𝑚(𝑛−𝑚) ]
′ 𝑥𝑛
It is transformed into a set of linear equations as follows:
′
𝑁𝑥1 = 𝑁𝑏1′ − (𝑁𝑑11 ′
𝑥𝑚+1 + 𝑁𝑑12 ′
𝑥𝑚+2 + ⋯ + 𝑁𝑑1(𝑛−𝑚) 𝑥𝑛 )
′ ′ ′ ′
𝑁𝑥2 = 𝑁𝑏2 − (𝑁𝑑21 𝑥𝑚+1 + 𝑁𝑑22 𝑥𝑚+2 + ⋯ + 𝑁𝑑2(𝑛−𝑚) 𝑥𝑛 )
………………………………………………………
′ ′ ′ ′
𝑁𝑥𝑚 = 𝑁𝑏𝑚 − (𝑁𝑑𝑚1 𝑥𝑚+1 + 𝑁𝑑𝑚2 𝑥𝑚+2 + ⋯ + 𝑁𝑑𝑚(𝑛−𝑚) 𝑥𝑛 )
This means that we were able to calculate 𝑚 in terms of (𝑛 − 𝑚)in terms of (𝑛 − 𝑚), 𝑥𝑚+1 , 𝑥𝑚+2 , … , 𝑥𝑛 ,we note
that the values of the variables 𝑥1 , 𝑥2 , … , 𝑥𝑚 , it relates to the values taken by the variables 𝑥𝑚+1 , 𝑥𝑚+2 , … , 𝑥𝑛 , or
in other words, what we give to the variables 𝑥𝑚+1 , 𝑥𝑚+2 , … , 𝑥𝑛 , and that for every sentence of Values such as
𝛽𝑚+1 , 𝛽𝑚+2 , … , 𝛽𝑛 for these variables we get a set of values for the variables 𝑥1 , 𝑥2 , … , 𝑥𝑚 is:
′
𝑁𝑥1 = 𝑁𝑏1′ − (𝑁𝑑11 ′
𝛽𝑚+1 + 𝑁𝑑12 ′
𝛽𝑚+2 + ⋯ + 𝑁𝑑1(𝑛−𝑚) 𝛽𝑛 )
′ ′ ′ ′
𝑁𝑥2 = 𝑁𝑏2 − (𝑁𝑑21 𝛽𝑚+1 + 𝑁𝑑22 𝛽𝑚+2 + ⋯ + 𝑁𝑑2(𝑛−𝑚) 𝛽𝑛 )
……………………………………………………………………
′ ′ ′ ′
𝑁𝑥𝑚 = 𝑁𝑏𝑚 − (𝑁𝑑𝑚1 𝛽𝑚+1 + 𝑁𝑑𝑚2 𝛽𝑚+2 + ⋯ + 𝑁𝑑𝑚(𝑛−𝑚) 𝛽𝑛 )
Thus, we obtain a solution that includes all the variables of sentence (5)
The solution is arranged as follows:
(𝛽1 , 𝛽2 , … , 𝛽𝑚 , 𝛽𝑚+1 , 𝛽𝑚+2 , … , 𝛽𝑛 )
But since the variables 𝑥𝑚+1 , 𝑥𝑚+2 , … , 𝑥𝑛 can take an infinite number of qualitative values (even if they are
restricted by certain conditions), we obtain an infinite number of corresponding values for the variables
𝑥1 , 𝑥2 , … , 𝑥𝑚 .
Therefore, if |𝑁𝐶| ≠ 0, then the set of equations (5) has an infinite number of acceptable solutions of the form:
(𝑥1 , 𝑥2 , … , 𝑥𝑚 , 𝑥𝑚+1 , 𝑥𝑚+2 , … , 𝑥𝑛 )
Thus, we obtain a solution that includes all variables of the sentence, which is the ordered solution:
(𝛽1 , 𝛽2 , … , 𝛽𝑚 , 𝛽𝑚+1 , 𝛽𝑚+2 , … , 𝛽𝑛 )
3. Basic solutions
Since sentence (5) has an infinite number of acceptable solutions, we will try to limit ourselves to a limited number
of them by setting the variables 𝑥𝑚+1 , 𝑥𝑚+2 , … , 𝑥𝑛 equal to zero. Then sentence (9) takes the following form:
1 0 0…0 𝑥1 𝑁𝑏1′
0 1 0…0 𝑥 2 𝑁𝑏 ′
[ ].[ .. ] = [ 2] (11)
…………… ..
0 0 0…1 𝑥𝑚 𝑁𝑏𝑚 ′

From it we get:

𝑥1 = 𝑁𝑏1′ , 𝑥2 = 𝑁𝑏2′ , … , 𝑥𝑚 = 𝑁𝑏𝑚

′

So, the complete solution is:

(𝑁𝑏1′ , 𝑁𝑏2′ , … , 𝑁𝑏𝑚
′
, 0,0, … ,0)
We call this solution the basic solution because it is attributed to the rule with single normal vectors in the space
𝑅𝑚 as follows:
1 0 0
0 1 0
0 0 0
𝑒1 = 𝑒2 = … … 𝑒𝑚 =
0 0 0
.. .. ..
[0 ] [ 0] [𝑚]
The set of vectors 𝑒1 , 𝑒2 , … , 𝑒𝑚 form a rule because they are linearly independent, and the vector 𝑁𝐵′ can be
expressed in terms of it using the factorials 𝑥1 , 𝑥2 , … , 𝑥𝑚 as follows:
𝑁𝐵′ = 𝑒1 𝑥1 + 𝑒2 𝑥2 + ⋯ + 𝑒𝑚 𝑥𝑚
We call the variables𝑥1 , 𝑥2 , … , 𝑥𝑚 , basic variables and we call other variables 𝑥𝑚+1 , 𝑥𝑚+2 , … , 𝑥𝑛
free or non-basic variables because they take qualitative values.

20
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

The process of choosing the variables 𝑥1 , 𝑥2 , … , 𝑥𝑚 to be basic variables is a random process, as we can form other
basic solutions, knowing that the possibilities available to obtain basic solutions are:
𝑛!
𝐶𝑛𝑚 =
𝑚! (𝑛 − 𝑚)!
It is a finite number of infinite acceptable solutions.
Example 1:
Find the joint solution of the following two linear equations:
2𝑥1 + 7𝑥2 + 3𝑥3 + 2𝑥4 = [2,5]
3𝑥1 + 9𝑥2 + 4𝑥3 + 𝑥4 = [3,7]
𝑥1 + 5𝑥2 + 3𝑥3 + 4𝑥4 = [4,8]
In the set of equations, the number of variables is 𝑛 = 4 and the number of equations is 𝑚 = 3. Therefore, the
number of basic variables is equal to 3 and the number of non-basic free variables is 𝑛 − 𝑚 = 1. The number of
possible solutions is calculated from the relationship:
𝑛!
𝐶𝑛𝑚 =
𝑚! (𝑛 − 𝑚)!
i.e.,
4!
𝐶43 = =4
3! (4 − 3)!
Write as follows:
(𝑥1 , 𝑥2 , 𝑥3 , 0) , (𝑥1 , 𝑥2 , 0, 𝑥4 ) , (𝑥1 , 0, 𝑥3 , 𝑥4 ), (0, 𝑥2 , 𝑥3 , 𝑥4 )
To obtain these solutions, we write the systems of equations in the following form:
2𝑥1 + 7𝑥2 + 3𝑥3 = [2,5] − 2𝑥4
3𝑥1 + 9𝑥2 + 4𝑥3 + 𝑥4 = [3,7] − 𝑥4
𝑥1 + 5𝑥2 + 3𝑥3 = [4,8] − 4𝑥4
The previous sentence is written in the following matrix form:
2 7 3 𝑥1 [2,5] 2
[3 9 4] . [𝑥2 ] = [[3,7]] − [1] . [𝑥4 ] (∗)
1 5 3 𝑥3 [4,8] 4
2 7 3 𝑥 1 [2,5] 2
𝐶 = [3 9 4] 𝑋 ′ = [𝑥2 ] 𝑁𝐵 = [[3,7]] 𝐷 = [1] 𝑋 " = [𝑥4 ]
1 5 3 𝑥 3 [4,8] 4
We calculate the determinant |𝐶| .
We find:
2 7 3
|𝐶| = |3 9 4| = −3 ≠ 0
1 5 3
2 7 3
To find the solutions, we find the reciprocal of the matrix, 𝐶 = [3 9 4]
1 5 3
We find:
−7 −1
2
3 3
𝐶 −1 = 5 −1
−1
3 3
[−2 1 1]
We compensate in the relationship:
𝑁𝐶 −1 . 𝑁𝐶. 𝑋 ′ = 𝑁𝐶 −1 . (𝑁𝐵 − 𝑁𝐷. 𝑋 " )
We get:

−7 −1 −7 −1
2 2 [2,5]
3 3 2 7 3 3 3 2
5 −1 . [3 9 4] = 5 −1 . ([[3,7]] − [1] . [𝑥4 ])
−1 1 5 3 −1 [4,8] 4
3 3 3 3
[−2 1 1] [−2 1 1]
−7 −1 −7 −1
2 [2,5] 2
1 0 0 𝑁𝑥1 3 3 3 3 2
[0 1 0] [𝑁𝑥2 ] = 5 −1 . [[3,7]] − 5 −1 . [1] . [𝑥4 ]
0 0 1 𝑁𝑥3 −1 [4,8] −1 4
3 3 3 3
[−2 1 1] [−2 1 1]
21
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

It is transformed into the following systems of equations:

−1
[0, ] −16
3
𝑁𝑥1 = 4 − [ 3 ] . [𝑥4 ]
− [1, ] 1
3
1
[ [3,5] ]
Setting the free variable 𝑥4 equal to zero we get:
−1
[0, ]
3
𝑁𝑥1 = 4
− [1, ]
3
[ [3,5] ]
i.e.,
−1 4
𝑥1 = [0, ] , 𝑥2 = − [1, ] , 𝑥3 = [3,5]
3 3
Thus, we obtain the first neutrosophic basic solution, which is:
−1 4
(𝑥1 , 𝑥2 , 𝑥3 , 0) = ([0, ] , − [1, ] , [3,5], 0)
3 3
In the same way we obtain other basic solutions.

4. Dissolved basic solutions:

The base solution is a degenerate and invalid solution if in the final result we obtain a value of zero for the variables
that we chose as the base.
Gaussian- Jordan method for solving a set of linear equations in which 𝒎 < 𝒏:
Based on the previous mathematical principles, the basic steps of the Gaussian- Jordan method are as follows:
1- We write the systems of equations (5) in the following matrix form:
𝐼. 𝑋 ′ + 𝑁𝐶 −1 . 𝐷. 𝑋 " = 𝑁𝐶 −1 . 𝑁𝐵 = 𝑁𝐵′
′
[𝐼 , 𝑁𝐶 −1 . 𝑁𝐷] . [𝑋 " ] = 𝑁𝐵′ (12)
𝑋
Which is written in the following detailed form:
′ ′ ′
1 0 0 … 0 𝑁𝑑11 𝑁𝑑12 … 𝑁𝑑1(𝑛−𝑚) 𝑥1 𝑁𝑏1′
0 1 0 … 0 𝑁𝑑21 𝑁𝑑22 … 𝑁𝑑1(𝑛−𝑚) 𝑥2
′ ′ ′
𝑁𝑏 ′
. [ . . ] = [ 2 ] (13)
……………………………………………… ..
[ 0 0 0 … 1 𝑁𝑑 ′
𝑚1 𝑁𝑑 ′
𝑚2 … 𝑁𝑑 ′
𝑚(𝑛−𝑚) ]
𝑥 𝑛 𝑁𝑏 𝑚
′

The transition from Figure (5) to Figure (12) is done by following the same steps that we mentioned in the previous
paragraph, but this method does not give us a basic solution unless we set all the free variables equal to zero. If we
do that, we only get the first solution, and to get all the solutions, we do the following steps:
a. We organize the following table:
Table No. (3) The first table for the Gaussian- Jordan method
Variables 𝑥1 𝑥2 …. 𝑥𝑚 𝑥𝑚+1 𝑥𝑚+2 …. 𝑥𝑛 𝑁𝐵
Equations
1 𝑎11 𝑎12 …. 𝑎1𝑚 𝑎1𝑚+1 𝑎1𝑚+2 …. 𝑎1𝑛 𝑁𝑏1′
2 𝑎21 𝑎22 … 𝑎2𝑚 𝑎2𝑚+1 𝑎2𝑚+2 …. 𝑎2𝑛 𝑁𝑏2′
…. … … … … … … … … …
′
𝑚 𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑚 𝑎𝑚𝑚+1 𝑎𝑚𝑚+2 … 𝑎𝑚𝑛 𝑁𝑏𝑚

b. We find the identity matrix 𝐼𝑚×𝑚 by processing the rows of the previous table in the same way that was
explained in the previous paragraph. This is done by specifying the variables that will be entered into the
base and let them be 𝑥1 , 𝑥2 , … , 𝑥𝑚 . As a result of this processing, we get the following table:

Table No. (4): Table of the first basic solution

Variables 𝑥1 𝑥2 …. 𝑥𝑚 𝑥𝑚+1 𝑥𝑚+2 …. 𝑥𝑛 𝑁𝐵′
Equations
′ ′ ′
1 1 0 …. 0 𝑁𝑑11 𝑁𝑑12 …. 𝑁𝑑1𝑛−𝑚 𝑁𝑏1′
′ ′ ′
2 0 1 … 0 𝑁𝑑21 𝑁𝑑22 …. 𝑁𝑑2𝑛−𝑚 𝑁𝑏2′
…. … … … … … … … … …
′ ′ ′
𝑚 0 0 … 1 𝑁𝑑𝑚1 𝑁𝑑22 … 𝑁𝑑𝑚𝑛−𝑚 𝑁𝑏2′
22
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

c. Setting all the free variables in Table (4) equal to zero, we obtain the following first basic solution:
(𝑁𝑏1′ , 𝑁𝑏2′ , … , 𝑁𝑏𝑚 ′
, 0,0, … ,0)
d. To obtain a second basic solution, we replace one of the basic variables, say 𝑥𝑚 , with one of the non-
′
basic variables 𝑥𝑚+1 , by selecting the appropriate pivot element, and here it is 𝑁𝑞𝑚1 . We work to delete
𝑥𝑚+1 from all equations except equation 𝑚 .We set the coefficient of this variable in this equation equal
to one. We perform the appropriate calculations through the two relations (4). We solve the following
second basic solution:
(𝑁𝑏1′ , 𝑁𝑏2′ , … , 𝑁𝑏𝑚−1
′ ′
, 0, 𝑁𝑏𝑚+1 , 0, … ,0)
To obtain other basic solutions, we repeat what was stated in step (d).
Non-negative basic solutions:
If all or some of the variables are conditioned to be non-negative, then some of the basic solutions are unacceptable
because they violate the condition. In such a case, we have to look for positive basic solutions from among the
basic solutions.
Due to the difficulty of applying the method mentioned in the example, especially in the case that includes a large
number of variables, the Gaussian- Jordan method was developed so that positive solutions are directly obtained.
The new method was called the simplex method, which is carried out according to the following steps:
The simplex method for finding non-negative basic solutions to a system of linear equations in which 𝒎 <
𝒏:
In the systems of equations (5):
1- We make all elements of the constant’s column 𝑁𝐵 on the second side of the equations non-negative,
by multiplying the equation whose second side is negative by (-1)
2- We put the coefficients of the new systems in a table.
3- We form a rule consisting of 𝑚 variables by choosing the variable that we want to enter into the rule,
for example, 𝑥𝑠 , and then we calculate the index.
𝑁𝑏𝑖 𝑁𝑏𝑡
𝜃 = 𝑀𝑖𝑛 [ ]= > 0; 𝑁𝑎𝑖𝑠 > 0, 𝑁𝑏𝑖 > 0
𝑁𝑎𝑖𝑠 𝑁𝑎𝑡𝑠
We call the element 𝑁𝑎𝑡𝑠 the pivot element, we delete the variable 𝑥𝑠 from all equations according
to the Gaussian- Jordan method, except for the equation 𝑡, in which its coefficient is equal to one.
We repeat the previous step until we form a base consisting of 𝑚 variables.
4- Setting the non-basic variables equal to zero we obtain the following non-negative basic solution:
(𝑁𝑏1′ , 𝑁𝑏2′ , … , 𝑁𝑏𝑚−1′ ′
, 0, 𝑁𝑏𝑚+1 , 0, … ,0)
5- To obtain other new non-negative basic solutions, we choose one of the variables to be a basic
variable, then we determine the pivot element and repeat the work we did for the previously
mentioned variable 𝑥𝑠 . We obtain a new non-negative basic solution, and so we continue working
until we obtain all non-negative basic solutions.

We explain the above through the following example:

Example 2:
𝑥1 − 3𝑥4 + 2𝑥5 = −[1,3]
𝑥2 + 2𝑥4 − 3𝑥5 = [2,8]

We multiply the first equation by (-1) until the condition 𝑁𝑏𝑖 > 0 is met, we obtain the following new systems:
−𝑥1 − 3𝑥3 − 2𝑥5 = [1,3]
𝑥2 + 2𝑥4 − 3𝑥5 = [2,8]
The stopping criterion is when we do not find a free column that we did not use for switching that contains a
positive element (at least one), meaning that all the elements of the free columns that were not used during the
swap are negative values.
In the systems of equations, the number of variables is 𝑛 = 5 and the number of equations is
𝑚 = 2 Therefore, the number of basic variables is equal to 2 and the number of non-basic free variables is 𝑛 −
𝑚 = 3. The number of possible solutions is calculated from the relationship:
𝑛!
𝐶𝑛𝑚 =
𝑚! (𝑛 − 𝑚)!
i.e.,
5!
𝐶52 = = 10
2! (5 − 2)!
Write as follows:
(𝑥1 , 𝑥2 , 0,0,0) , (𝑥1 , 0, 𝑥3 , 0,0), (𝑥3 , 0,0, 𝑥4 , 0), (𝑥1 , 0,0,0, 𝑥5 ), (0, 𝑥2 , 𝑥3 , 0,0)
(0, 𝑥2 , 0, 𝑥4 , 0), (0, 𝑥2 , 0,0, 𝑥5 ), (0,0, 𝑥3 , 𝑥4 , 0), (0,0, 𝑥3 , 0, 𝑥5 ), (0,0,0, 𝑥4 , 𝑥5 )

23
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

To obtain these solutions, we organize the following table:

Table No. (5): The first table for the simplex method
Variables 𝑥1 𝑥2 𝑥3 𝑥4 𝑥5 𝑁𝐵
Equations
1 −1 0 0 3 −2 [1,3]
2 0 1 0 2 −3 [2,8]

To find a basic solution to the set of equations, we choose a variable, for example 𝑥4 , to be a basic variable, and
to determine the appropriate anchor element, we calculate the index:
𝑁𝑏𝑖 [1,3] [2,8] [1,3]
𝜃 = 𝑀𝑖𝑛 [ ] = 𝑀𝑖𝑛 [ , ]=
𝑁𝑎𝑖𝑠 3 2 3
That is, the fulcrum is 𝑎14 = 3. By performing the necessary calculations to delete the variable 𝑥4 from the two
equations, we obtain the following table:
Table No. (6) The second table for the simplex method
Variables 𝑥1 𝑥2 𝑥3 𝑥4 𝑥5 𝑁𝐵′
Equations
𝑥4 1 0 0 1 2 1
− − [ , 1]
3 3 3
2 2 1 0 0 5 4
[ , 6]
3 3 3

We choose another variable to be a basic variable. We note that the variable 𝑥2 is ready to be a basic variable,
and thus we get the following table:
Table No. (7): Final solution table
Variables 𝑥1 𝑥2 𝑥3 𝑥4 𝑥5 𝑁𝐵′
Equations
𝑥4 1 0 0 1 2 1
− − [ , 1]
3 3 3
𝑥2 2 1 0 0 5 4
[ , 6]
3 3 3

Thus, we obtain a base consisting of the variables 𝑥2 , 𝑥4 . We set the free variables equal to zero, and we obtain
the following non-negative neutrosophic basic solution:
4 1
(0, [ , 6] , 0, [ , 1] , 0)
3 3
To obtain other solutions, we repeat what we did to determine the previous solution.

5. Conclusion and results

As a basis for neutrosophic linear programming, we presented in this research a study of the sets of neutrosophic
linear equations, and the Gaussian- Jordan method, which is considered the mathematical basis for the simplex
method used to find positive basic solutions, which can be used when there are restrictions on some or all of the
variables to be positive values, which in turn was the basis for the method. Direct simplex used to find the optimal
solution for linear models. Through the examples that we presented on the systems of neutrosophic equations, we
arrived at basic neutrosophic solutions that express unspecified values. Such sentences can be used in cases where
the data provided to the systems that operate according to these systems of equations is subject to change. Here
we can benefit from the margin of freedom offered by neutrosophic values.

References
[1] Florentin Smarandache, Maissam Jdid, On Overview of Neutrosophic and Plithogenic Theories and
Applications, Applied Mathematics and Data Analysis, Vo .2, No .1, 2023
[2] Mountajab Alhasan, Malath F Alaswad. A Study of the Movement of a Neutrosophic Material Point in
the Neutrosophic Plane by Using a Neutrosophic AH-Isometry, Vol 2, No 1, 2023
[3] Wang, J., & Zhang, H. (2017). Generalized Neutrosophic Sets and Their Application in Multi-criteria
Decision-making. Springer.
[4] Ye, J. (2019). Neutrosophic logic and its applications in information fusion. Information Fusion, 48, 32-
40.
[5] Maissam Jdid, Math 1, Faculty of Informatics Engineering, Al-Sham Private University
Publications,2020. (Arabic version)

24
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023
International Journal of Neutrosophic Science (IJNS) Vol. 23, No. 02, PP. 16-25, 2024

[6] Maissam Jdid, Operations Research, Faculty of Informatics Engineering, Al-Sham Private University
Publications,2021. (Arabic version)
[7] Alali. Ibrahim Muhammad, Operations Research. Tishreen University Publications, 2004. (Arabic
version)
[8] Bugaha J.S, Mualla.W, and others -Operations Research Translator into Arabic, The Arab Center for
Arabization, Translation, Authoring and Publishing, Damascus,1998. (Arabic version).

25
Doi: https://doi.org/10.54216/IJNS.230202
Received: June 12, 2023 Revised: October 18, 2023 Accepted: November 14, 2023