Dual Problem of an LPP

Given a LPP (called the primal problem), we shall associate another LPP called the dual problem of the original (primal) problem. We shall see that the Optimal values of the primal and dual are the same provided both have finite feasible solutions. This topic is further used to develop another method of solving LPPs and is also used in the sensitivity (or post-optimal) analysis.

Definition of the dual problem Given the primal problem (in standard form) Maximize z c1 x1 c2 x2 ... cn xn subject to
a11 x1 a12 x2 ... a1n xn b1 a21 x1 a22 x2 ... a2 n xn b2 . . am1 x1 am 2 x2 ... amn xn bm x1 , x2 ,..., xn 0, b1 , b2 ,..., bm 0

the dual problem is the LPP

Minimize subject to
w b1 y1 b2 y2 ... bm ym

a11 y1 a21 y2 ... am1 ym c1 a12 y1 a22 y2 ... am 2 ym c2 . . a1n y1 a2 n y2 ... amn ym cn y1 , y2 ,..., yn unrestricted in sign

If the primal problem (in standard form) is

Minimize z c1 x1 c2 x2 ... cn xn subject to

a11 x1 a12 x2 ... a1n xn b1 a21 x1 a22 x2 ... a2 n xn b2

. . am1 x1 am 2 x2 ... amn xn bm x1 , x2 ,..., xn 0, b1 , b2 ,..., bm 0

Then the dual problem is the LPP

Maximize w b1 y1 b2 y2 ... bm ym

subject to
a11 y1 a21 y2 ... am1 ym c1 a12 y1 a22 y2 ... am 2 ym c2 . . a1n y1 a2 n y2 ... amn ym cn y1 , y2 ,..., yn unrestricted in sign

We thus note the following: 1. In the dual, there are as many (decision) variables as there are constraints in the primal. We usually say yi is the dual variable associated with the ith constraint of the primal.

2. There are as many constraints in the dual as there are variables in the primal.

3. If the primal is maximization then the dual is minimization and all constraints are If the primal is minimization then the dual is maximization and all constraints are 4. In the primal, all variables are 0 while in the dual all the variables are unrestricted in sign.

5. The objective function coefficients cj of the primal are the RHS constants of the dual constraints. 6. The RHS constants bi of the primal constraints are the objective function coefficients of the dual.

7. The coefficient matrix of the constraints of the dual is the transpose of the coefficient matrix of the constraints of the primal.

Problem 4(a) Problem Set 4.1A Page 119

Write the dual of the LPP Maximize subject to

z 5 x1 2 x2

x1 x2 2 2 x1 3 x2 5 x1 , x2 0

Thus the primal in the standard form is: Maximize z 5 x1 2 x2 0 x3 0 x4 subject to

x1 x2 x3 2 x1 3 x2 x1 , x2 , x3 , x4 0

2 x4 5

Hence the dual is:

Minimize subject to

w 2 y1 5 y2

y1 2 y2 5 y1 3 y2 y1 y2 2 0 0

y1 0, y2 0

y1 , y2 unrestricted in sign

Problem 4(b) Problem Set 4.1A Page 119

Write the dual of the LPP Minimize z 6 x1 3x2 subject to

6 x1 3 x2 x3 2 x1 , x2 , x3 0

3 x1 4 x2 x3 5

Thus the primal in the standard form is: Minimize z 6 x1 3x2 0 x3 0 x4 0 x5

subject to

6 x1 3 x2 x3 x4 3 x1 4 x2 x3 x1 , x2 , x3 , x4 , x5 0

2 x5 5

Hence the dual is:

Maximize w 2 y1 5 y2 subject to
6 y1 3 y2 6 3 y1 4 y2 3 y1 y1 y2 0 0 y2 0 y1 , y2 unrestricted in sign

y1 , y2 0

Problem 4(c) Problem Set 4.1A Page 119

Write the dual of the LPP Maximize subject to

z x1 x2
2 x1 x2 5 3x1 x2 6 x1 , x2 unrestricted in sign

Thus the primal in the standard form is: Maximize subject to

zx x x x
1 1 2

2
2

2x 2x x x 5 3x 3x x x 6 x ,x ,x ,x 0
1 1 2 2 1 1 2 2

Hence the dual is:

Minimize subject to
2 y1 3 y2 1 2 y1 3 y2 1 y1 y1 y2 1 y2 1

w 5 y1 6 y2
2 y1 3 y2 1

y1 y2 1

y1 , y2 unrestricted in sign

From the above examples we get the following SOB rules for writing the dual: Label Sensible Odd Bizarre Sensible Odd Bizarre Maximization Constraints form = form form Variables 0 unrestricted 0 Minimization Variables 0 unrestricted 0 Constraints form = form form

Theorem: The dual of the dual is the primal (original problem).

Proof. Consider the primal problem (in standard form) z c1 x1 c2 x2 ... cn xn Maximize subject to
a11 x1 a12 x2 ... a1n xn b1 a21 x1 a22 x2 ... a2 n xn b2 . . am1 x1 am 2 x2 ... amn xn bm x1 , x2 ,..., xn 0

The dual problem is the LPP

Minimize w b1 y1 b2 y2 ... bm ym

subject to

a11 y1 a21 y2 ... am1 ym c1 a12 y1 a22 y2 ... am 2 ym c2 . . a1n y1 a2 n y2 ... amn ym cn y1 , y2 ,..., yn unrestricted in sign

Case (i): All cj 0. Then the dual problem in the standard form is the LPP
Minimize subject to
w b1 y b1 y ... bm y bm y 0t1 0t 2 ... 0t n
1 1 m m

a11 y1 a11 y1 ... am1 ym am1 ym t1 c1 a12 y1 a12 y1 ... am 2 ym am 2 y m t 2 c2

. .
a1n y1 a1n y1 ... amn ym amn ym t n cn y1 , y1 ,..., ym , ym , t1 , t 2 ,...,t n 0

Hence its dual is the LPP Maximize z c1 x1 c2 x2

subject to

... cn xn

a11 x1 a12 x2 ... a1n xn b1 a11 x1 a12 x2 ... a1n xn b1 . am1 x1 am 2 x2 ... amn xn bm am1 x1 am 2 x2 ... amn xn bm x1 0, x2 0,..., xn 0 x1 , x2 ,..., xn unrestricted in sign

Which is nothing but the LPP

Maximize z c1 x1 c2 x2 ... cn xn subject to
a11 x1 a12 x2 ... a1n xn b1 a21 x1 a22 x2 ... a2 n xn b2 . . am1 x1 am 2 x2 ... amn xn bm x1 , x2 ,...,xn 0

This is the primal problem.

Case (ii): All cj 0 except c1. Then the dual problem in the standard form is the LPP Minimize
w b1 y1 b1 y1 ... bm ym bm ym

Subj ect a11 y1 a11 y1 ... am1 ym am1 ym t1 c1 a12 y1 a12 y1 ... am 2 ym am 2 ym t 2 c2 to . .
a y a y ... amn y amn y t n cn y , y ,..., y , y , t1 , t 2 ,...,t n 0
1 1 m m 1n 1 1n 1 m m

0t1 0t 2 ... 0t n

Hence its dual is the LPP Maximize z c1v1 c2 x2 ... cn xn

subject to
a11v1 a12 x2 ... a1n xn b1 a11v1 a12 x2 ... a1n xn b1 . am1v1 am 2 x2 ... amn xn bm am1v1 am 2 x2 ... amn xn bm v1 0, x2 0,..., xn 0 v1 , x2 ,..., xn unrestricted in sign

Which is nothing but the LPP

Maximize

z c1v1 c2 x2 ... cn xn
a11v1 a12 x2 ... a1n xn b1 a21v1 a22 x2 ... a2 n xn b2 . . am1v1 am 2 x2 ... amn xn bm v1 0, x2 ,...,xn 0

subject to

Putting x1 v1

This is nothing but the LPP

Maximize subject to
z c1 x1 c2 x2 ... cn xn
a11 x1 a12 x2 ... a1n xn b1 a21 x1 a22 x2 ... a2 n xn b2 . . am1 x1 am 2 x2 ... amn xn bm x1 , x2 ,..., xn 0

This is the primal problem. Cases where other cj are 0 are similarly treated.