Some Problems Illustrating The Principles of Duality

Some problems illustrating the
principles of duality
In this lecture we look at some
problems that uses the results from
Duality theory
Problem 1
Consider the LPP

Maximize z  5 x1  2 x2  3 x3
subject to
x1  5 x2  3 x3  b1
x1  5 x2  6 x3  b2
x1 , x2 , x3  0
The following optimal tableau corresponds to

specific values of b1 and b2:
Basic z x1 x2 x3 x4 x5 Sol
z 1 0 a 7 d e 150
x1 0 1 b 2 1 0 30
x5 0 0 c -8 -1 1 10
Determine the following:

(a)The RHS values b1, b2
(b)The entries a, b, c, d, e
(c)The optimal solution of the dual.
(a) The optimal solution of the primal is B -1b, where
1 1 0  b1 
B    b
 1 1 b 
 2
Thus
 b1  30
B-1b =  b  b     Hence b1 = 30, b2 = 40
 1 2 10 
(b) b   B 1 A   1 0   5   5 
c  2  1 1   5  10 
      
Hence b = 5, c = -10
 5 
a = z2  c2  CB B A2  c2   5 0  10   2  23
1
 
1
d = z4  c4  CB B A4  c4   5 0    0 
1
5
 1
e = 0 as x5 is a basic variable.
(c) The optimal solution of the dual is

 1 0
 y1 y2   C B B   5 0  
1
   5 0
 1 1 
Note: The optimal solution of the dual is also
given by [ d, e] = [5, 0].
Problem 1
Estimate a range for the optimal objective value of
the following LPP
Minimize z  5 x1  2 x2
subject to x1  x2  3
2 x1  3 x2  5
x1 , x2  0
An obvious feasible solution is:
x1=3, x2=0 with z = 15. Thus the optimal value  15.
The dual of the given LPP is the LPP
Maximize w  3 y1  5 y 2
subject to y1  2 y2  5
 y1  3 y2  2
y1 , y2  0
An obvious feasible solution of the dual is:
y1=3, y2=1 with w = 14. Thus the optimal value of
the dual  14.
Hence by the weak duality theorem: 14  z* 15,
z* being the optimal value of z.
Problem 2
In Problem 1, let y1, y2 be the dual variables.
Determine whether the following pairs of primal-
dual solutions are optimal:
x1  3, x2  1; y1  4, y2  1
Though z = 17 = w, neither pair satisfy the

constraints of the respective problem. Hence not
even solutions.
Aliter: Since z* = w*  15, and since the given
pairs give z = 17 = w, they cannot be optimal.
We shall now do some problems from section 4.2.3.
The theme is: How to find the optimal solution of
the dual from the primal?
As we have seen in chapter 7, the dual solution is
given by Y  CBB 1, the so-called Simplex
multiplier.
In words, we can say
[Optimal values of dual variables]
= Row vector of original objective coefficients of
optimal primal variables *Optimal primal inverse.
Now we note that in optimal tableau, the
coefficient of xj in z-Row is zj - cj
1
 CB B A j  c j
 a1 j 
a 
 2j
  y1 y2 . . ym   .   cj
 
 . 
 amj 
 
= LHS of the jth dual constraint – RHS of the jth
dual constraint.
We can find the dual solution using this
equation. In your book this method of
finding the dual solution is referred to as
Method 2 while the earlier one is referred to
as Method 1.
Problem 3
Consider the LPP
Maximize z  x1  5 x2  3x3
subject to
x1  2 x2  x3  3
2 x1  x2 4
x1 , x2 , x3  0
(a) Write the associated dual problem.
(b) Given the optimal basic variables are x1 and x3,
determine the associated optimal dual solution.
(a)The dual problem is:
Minimize w  3 y1  4 y2
subject to y1  2 y2  1
2 y1  y2  5
y1 3
y1 , y2 unrestricted in sign
(b) 1 1  0 1/ 2 
B  ; 1
B   ; C B   1 3
 2 0 1 1/ 2 
Hence the optimal dual solution (by method 1) is:
0 1/ 2 
 y1 y2   CB B   1 3 
1

1 1/ 2 
  3 1
Problem 4
Consider the LPP
Maximize z  2 x1  4 x2  4 x3  3 x4
subject to x1  x2  x3 4
x1  4 x2  x4  8
x1 , x2 , x3 , x4  0
The primal objective row is given by
z  2 x1  0 x2  0 x3  3 x4  16
Use this information to find the associated optimal
dual solution.
We first write the associated dual problem.
Dual: Minimize w  4 y1  8 y2
subject to
y1  y2  2
y1  4 y2  4
y1  4
y2  3
z  2 x1  0 x2  0 x3  3 x4  16
Hence by method 2, the dual optimal solution
satisfy the equations:
coefficient of x1 in optimal row: 2  y1  y2  2
coefficient of x2 in optimal row: 0  y1  4 y2  4
coefficient of x3 in optimal row: 0  y1  4
coefficient of x4 in optimal row: 3  y2  ( 3)
From the last two equations we get y1  4, y2  0

Problem 5
Consider the LPP
Maximize z  5 x1  2 x2  3 x3
subject to x  5 x  2 x  30
1 2 3
x1  5 x2  6 x3  40
x1 , x2 , x3  0
z  0 x1  23x2  7 x3  (5  M ) x4  0 x5  150
where artificial x4 and slack x5 are the starting basic
Write the associated dual problem and determine
its optimal solution from the optimal z-equation.
Since clearly Big-M method has been used to solve
the primal, the primal (in the modified form) is:
Maximize z  5 x1  2 x2  3 x3  Mx4  0 x5
subject to x1  5 x2  2 x3  x4  30
x1  5 x2  6 x3  x5  40
x1 , x2 , x3 , x4 , x5  0
Hence the dual problem is:
Minimize w  30 y1  40 y2
subject to
y1  y2  5
5 y1  5 y2  2
2 y1  6 y2  3
y1  M
y2  0
z  0 x1  23x2  7 x3  (5  M ) x4  0 x5  150
Hence by method 2, the dual optimal solution
satisfy the equations:
coefficient of x4 in optimal row: 5  M  y1  ( M )
coefficient of x1 in optimal row: 0  y1  y2  5
Hence we get y1  5, y2  0

Some Problems Illustrating The Principles of Duality

Uploaded by

Copyright:

Available Formats

Some Problems Illustrating The Principles of Duality

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Some Problems Illustrating The Principles of Duality

Uploaded by

Copyright:

Available Formats

Some problems illustrating the

Consider the LPP

The following optimal tableau corresponds to

Determine the following:

(c) The optimal solution of the dual is

Though z = 17 = w, neither pair satisfy the

From the last two equations we get y1  4, y2  0

You might also like