0% found this document useful (0 votes)
245 views7 pages

Dual Simplex Method

The document describes the steps of the dual simplex method to solve linear programming problems (LPP). The key steps are: 1) Express the LPP in standard form with non-negative variables and an identity submatrix. 2) Express the objective function in terms of non-basic variables and ensure the coefficients have the correct sign based on maximization or minimization. 3) Create an initial dual simplex table with the initial basic feasible solution. 4) Select the leaving variable using minimum ratio test and pivot to find the entering variable. 5) Update the dual simplex table using row operations until an optimal solution is reached where all variables are non-negative. Two examples are

Uploaded by

math.master1230
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
245 views7 pages

Dual Simplex Method

The document describes the steps of the dual simplex method to solve linear programming problems (LPP). The key steps are: 1) Express the LPP in standard form with non-negative variables and an identity submatrix. 2) Express the objective function in terms of non-basic variables and ensure the coefficients have the correct sign based on maximization or minimization. 3) Create an initial dual simplex table with the initial basic feasible solution. 4) Select the leaving variable using minimum ratio test and pivot to find the entering variable. 5) Update the dual simplex table using row operations until an optimal solution is reached where all variables are non-negative. Two examples are

Uploaded by

math.master1230
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Dual Simplex Method

■ Procedure to solve the LPP using Dual simplex method:


Consider the LPP:
Maximize (or Minimize) f (x) = cT x

23
subject to
Ax ≤ (or ≥) b
x ≥ 0

20
where c = {c1 , c2 , . . . , cn }T , x = {x1 , x2 , . . . , xn }T , b = {b1 , b2 , . . . , bn }T and A = [aij ]m×n .

-
22
The dual simplex method is applicable if the optimality condition is satisfied in the above format.
Following are the steps used in the Dual simplex method.

20
Step 1: Express the LPP in the following format:

Min. or Max. f (x) = cT x

r
subject to

te
Ax = b
x ≥ 0

es
where one or more bi ’s are negative and A has a identity matrix as a submatrix.
m
Step 2: Express f (x) in terms of non-basic variables.
{The cost coefficients of all non-basic variable after expressing f (x) in terms of non-basic variables only must be all
Se

negative (positive) for maximization (minimization) type problem}.

Step 3: The above gives a optimal basic (but not feasible) solution xB = (x1 , x2 , . . . , xm ). Make the initial dual simplex
dd

table.
Step 4: In dual simplex method, the leaving variable is decided first. A basic variable xr is selected for laving the basis using
:O

the following:
xr = min {xi , xi < 0}
1≤i≤m
5

Step 5: After selecting the leaving variable xr , the entering variable xk is selected as follows:
51

 
zk − ck zj − cj j
θ= = min ; αr < 0
αrk j αrj
A-

If all αrj are non-negative, the problem has no feasible solution.


M

Step 6: The next dual simplex table is obtained using the row operations with the help of the pivot row and pivot element.
(same procedure as in the regular simplex method).
Step 7: When a basic solution reached, in which all variables are non-negative and We obtain the optimal solution.

1
■ Examples:
1. Solve the following LPP using Dual Simplex method.

Minimize z = 3x1 + x2
Subject to

23
x1 + x2 ≥ 1
2x1 + 3x2 ≥ 2
x1 , x2 ≥ 0

20
Converting the given constraints to ≤ type and introducing slack variables s1 and s2 in first and second constraints

-
respectively.

22
The modified LPP is:

Minimize z = 3x1 + x2

20
Subject to
−x1 − x2 + s1 = −1
−2x1 − 3x2 + s2 = −2

r
x1 , x2 , s1 , s2 ≥ 0

te
The starting basic solution (s1 , s2 ) = (−1, −2) does not satisfy the non-negativity. The corresponding table is:

Basic x1

x2 y

es s1 s2 Solution
m
z −3 −1 0 0 0
Se

s1 −1 −1 1 0 −1

←− s2 −2 -3 0 1 -2
dd

Since the above table satisfies optimality condition but does not satisfy the feasibility, Dual simplex method is appli-
:O

cable.

The basic variable having most negative value (= −2) in the above table leaves the basis and the minimum ratio
indicates that x2 enters the basis. The next table is:
5
51


Basic x1 x2 s1 s2 y Solution

7 1 2
A-

z - 0 0 -
3 3 3
1 1 1
←− s1 - 0 1 - -
M

3 3 3
2 1 2
x2 1 0 -
3 3 3

The previous table indicates that s1 and s2 are the leaving and entering variables, respectively. Thus, the next table
is:

Basic x1 x2 s1 s2 Solution
z −2 0 −1 0 1
s2 1 0 −3 1 1

x2 1 1 −1 0 1

Since all basic variables are non-positive, last table is the optimal feasible table for the modified problem. The opti-
mal feasible solution for the modified problem is x1 = 0, x2 = 1 and f (x) = 1.

2
2. Solve the following LPP using Dual Simplex method.

Maximize z = −7x1 − 3x2


Subject to
5x1 − 2x2 ≥ 12
6x1 + 3x2 ≤ 10

23
x1 , x2 ≥ 0

20
Converting the first constraint to ≤ type and introducing slack variables s1 and s2 in first and second constraints
respectively.
The modified LPP is:

-
22
Maximize z = −7x1 − 3x2
Subject to
−5x1 + 2x2 + s1 = −12

20
6x1 + 3x2 + s2 = 10
x1 , x2 , s1 , s2 ≥ 0
The starting basic solution (s1 , s2 ) = (−12, 10) does not satisfy the non-negativity. The corresponding table is:

r
te

Basic x1 y x2 s1 s2 Solution

←−
z
s1
7
-5 es
3
2
0
1
0
0
0
-12
m
s2 6 3 0 1 10
Se

Since the above table satisfies optimality condition but does not satisfy the feasibility, Dual simplex method is appli-
cable.
dd

The basic variable having most negative value (= −12) in the above table leaves the basis and the minimum ratio
indicates that x1 enters the basis. The next table is:
:O

Basic x1 x2 s1 s2 Solution
29 7 84
5

z 0 0 -
5 5 5
51

2 1 12
x1 1 - - 0
5 5 5
27 6 22
←− s2 0 1 -
A-

5 5 5

The previous table indicates that s2 is the leaving variable. Since all the coefficients are non-negative in the s2 row,
M

there is no entering variable. This indicates that the problem has no feasible solution.
■ Addition of a Constraint Examples
1. The following LPP is solved using the Dual Simplex method (or any other suitable method).

Minimize f (x) = 3x1 + x2


Subject to
x1 + x2 ≥ 1
2x1 + 3x2 ≥ 2
x1 , x2 ≥ 0

If we add one more constraints


3x1 + 2x2 ≥ 2
The modified problem is:
Minimize f (x) = 3x1 + x2
Subject to
x1 + x2 ≥ 1
2x1 + 3x2 ≥ 2
3x1 + 2x2 ≥ 2
x1 , x2 ≥ 0

3
Do we need to solve the above modified problem again to get the optimal solutions?
We can obtain the optimal solution for the modified problem (without solving again) using the optimal solution of
the original problem.

The optimal table for the given problem is:

23
Basic x1 x2 s1 s2 Solution
f (x) −2 0 −1 0 1

20
s2 1 0 −3 1 1

-
x2 1 1 −1 0 1

22
After adding new constraint in the optimal table of the original problem, we get the following table:

20
Basic x1 x2 s1 s2 s3 Solution
f (x) −2 0 −1 0 0 1

r
te
s2 1 0 −3 1 0 1

x2 1 1 −1 0 0 1

s3 −3 −2 es
−1 0 1 −2
m
Se

To make f (x) − row consistent before proceeding further, the coefficients of s3 (being the basic variables) needs to
be made zero in the f (x) − row. Thus, the corresponding simplex table is:
dd

Basic x1 x2 s1 s2 s3 Solution
f (x) −2 0 −1 0 0 1
:O

s2 1 0 −3 1 0 1

x2 1 1 −1 0 0 1
5

s3 −1 0 −3 0 1 0
51
A-

Since all basic variables are non-negative, last table is the optimal feasible table for the modified problem. The
optimal feasible solution for the modified problem is (x1 , x2 ) = (0, 1) with M in f (x) = 1.
M

2. Solve the following LPP using Dual Simplex method (or any other suitable method).

Minimize z = 2x1 + 3x2 + 4x3 + 5x4


Subject to
x1 − x2 + x3 − x4 ≥ 10
x1 − 2x2 + 3x3 − 4x4 ≥ 6
3x1 − 4x2 + 5x3 − 6x4 ≥ 15
x1 , x2 , x3 , x4 ≥ 0
After solving the above LPP add the constraint x1 + 2x2 + 3x3 − 4x4 ≤ 8 and find the optimal solution of the
modified LPP with the help of the optimal table of the original problem.
Converting the given constraints to ≤ type and introducing slack variables s1 , s2 and s3 .
The modified LPP is
Minimize z = 2x1 + 3x2 + 4x3 + 5x4 + 0s1 + 0s2 + 0s3
Subject to
−x1 + x2 − x3 + x4 + s1 = −10
−x1 + 2x2 − 3x3 + 4x4 + s2 = −6
−3x1 + 4x2 − 5x3 + 6x4 + s3 = −15
x1 , x2 , x3 , x4 , s1 , s2 , s3 ≥ 0

4

Basic x1 y x2 x3 x4 s1 s2 s3 Solution

z −2 −3 −4 −5 0 0 0 0
s1 −1 1 −1 1 1 0 0 −10

s2 −1 2 −3 4 0 1 0 −6

23
←− s3 -3 4 −5 6 0 0 1 -15

- 20
The starting basic solution (s1 , s2 , s3 ) = (−10, −6, −15) does not satisfy the non-negativity. The corresponding table

22
is:
The basic variable having most negative value (= −15) in the above table leaves the basis and the minimum ratio
indicates that x1 enters the basis. The next table is:

20

Basic x1 x2 x3 x4 s1 s2 s3 y Solution

r
17 2 2

te
z 0 - - −9 0 0 - 10
3 3 3
1 2 1

es
←− s1 0 - −1 1 0 - -5
3 3 3
2 4 1
s2 0 - 2 0 1 - −1
m
3 3 3
4 5 1
x1 1 - −2 0 0 - 5
Se

3 3 3

The previous table indicates that s1 and s3 are the leaving and entering variables, respectively. Thus, the next table
dd

is:

Basic x1 x2 x3 x4 s1 s2 s3 Solution
:O

z 0 −5 −2 −7 −2 0 0 20
s3 0 1 −2 3 −3 0 1 15
5

s2 0 1 −2 3 −1 1 0 4
51

x1 1 −1 1 −1 −1 0 0 10
A-

Since all basic variables are non-negative, last table is the optimal feasible table for the modified problem. The
M

optimal feasible solution for the modified problem is x1 = 10, x2 = 0, x3 = 0 and f (x) = 20.

The given constraint x1 + 2x2 + 3x3 − 4x4 ≤ 8 can be added to the above optimal table for the original problem. To
do so, first convert the given constraint to equality by adding a slack variable s4 .

x1 + 2x2 + 3x3 − 4x4 + s4 = 8

Afterwards, the optimal table for the original problem is modified by adding the variable s4 to the basis (and
corresponding constraint as a new row in the table) and a new column corresponding the slack variable s4 .
The modified table is:
The above table is not a Dual simplex table. To make above table consistent before proceeding further, the coeffi-
cients of x1 (being the basic variables) need to be made zero in the s4 − row. Thus, the next Dual simplex table
is:
Since the above table satisfies optimality condition but does not satisfy the feasibility, Dual simplex method is appli-
cable.

The basic variable having most negative value = −2 in the last table leaves the basis and the minimum ratio indicates
that x4 enters the basis. The next table is:

5
Basic x1 x2 x3 x4 s1 s2 s3 s4 Solution
z 0 −5 −2 −7 −2 0 0 0 20
s3 0 1 −2 3 −3 0 1 0 15

s2 0 1 −2 3 −1 1 0 0 4

23
x1 1 −1 1 −1 −1 0 0 0 10

20
s4 1 2 3 −4 0 0 0 1 8

-
22

Basic x1 x2 x3 x4 y s1 s2 s3 s4 Solution

z 0 −5 −2 −7 −2 0 0 0 20

20
s3 0 1 −2 3 −3 0 1 0 15

s2 0 1 −2 3 −1 1 0 0 4

r
te
x1 1 −1 1 −1 −1 0 0 0 10

←− s4 0 3 2
es
-3 1 0 0 1 -2
m
Basic x1 x2 x3 x4 s1 s2 s3 s4 Solution
Se

20 13 7 74
z 0 −12 - 0 - 0 0 -
3 3 3 3
s3 0 4 0 0 −2 0 1 1 13
dd

s2 0 4 0 0 0 1 0 1 4
:O

1 4 1 32
x1 1 −2 0 - 0 0 -
3 3 3 3
2 1 1 2
x4 0 −1 - 1 - 0 0 -
5

3 3 3 3
51

Since all basic variables are non-negative, last table is the optimal feasible table for the modified problem. The
32 2 74
A-

optimal feasible solution for the modified problem is x1 = , x2 = 0, x3 = 0, and x4 = with f (x) = .
3 3 3
M

Dual Simplex Method Practice Example

1. Solve the following LPP using Dual Simplex method.


Minimize z = 3x1 + 2x2 + x3
subject to
3x1 + x2 + x3 ≥ 3
−3x1 + 3x2 + x3 ≥ 6
x1 + x2 + x3 ≤ 3
x1 , x2 , x3 ≥ 0

2. Solve the following LPP using Dual Simplex method.


Maximize z = −3x1 − x2
subject to
x1 + x2 ≥ 1
2x1 + 3x2 ≥ 2
x1 , x2 , ≥ 0

6
3. Solve the following LPP using Dual Simplex method.

Minimize z = 20x1 + 16x2


subject to
x1 ≥ 5
x2 ≥ 6

23
2x1 + x2 ≥ 17
x1 + x2 ≥ 12
x1 , x2 , ≥ 0

20
4. Solve the following LPP using Dual Simplex method.

-
22
Maximize f (x) = −x1 − 2x2 − x3
subject to
x1 + 3x2 + x3 ≥ 17

20
3x1 + 2x2 − x3 ≤ 15
x1 − x2 + 2x3 ≤ 5
x1 , x2 , x3 ≥ 0

r
te
5. Solve the following LPP using Dual Simplex method.

es
Maximize f (x) = −2x1 − x2
subject to
m
2x1 + x2 ≤ 7
3x1 + 2x2 ≥ 16
Se

x1 + 2x2 ≥ 5
x1 , x2 ≥ 0
dd
5 :O
51
A-
M

You might also like