Scribd
Upload
35 views14 pages

Idea of A Dual

Management Science

Uploaded by

pavan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
35 views14 pages

Idea of A Dual

Management Science

Uploaded by

pavan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 14

Idea of a Dual

Maximize 500X1 + 600X2


3X1 + 3X2 ≤ 27 Without solving the problem, can you give an upper
2X1 + 4X2 ≤ 32 estimate of Z ?
X1, X2 ≥ 0

Multiply constraint 2 by 300 to get 600X1 + 1200X2 ≤ 9600


Multiply constraint 1 by 200 to get 600X1 + 600X2 ≤ 5400
Multiply constraint 1 by 100, Multiply constraint 2 by 100 and
add to get 500X1 + 700X2 ≤ 5900
If we multiply constraint 1 by Y1 and multiply constraint 2 by Y2
and add. If 3Y1 + 2Y2 ≥ 500 and 3Y1 + 4Y2 ≥ 600 and Y1, Y2 ≥ 0,
then 27Y1 + 32Y2 is an upper estimate of Z.

A good estimate gives minimum value of 27Y1 + 32Y2


Given problem (Primal) Dual

Maximize 500X1 + 600X2 Minimize 27Y1 + 32Y2


3X1 + 3X2 ≤ 27 3Y1 + 2Y2 ≥ 500
2X1 + 4X2 ≤ 32 3Y1 + 4Y2 ≥ 600
X1, X2 ≥ 0 Y1, Y2 ≥ 0
Solution is (2, 7) Z = 5200
Minimize 27Y1 + 32Y2
3Y1 + 2Y2 ≥ 500
Maximize 500X1 + 600X2 3Y1 + 4Y2 ≥ 600
3X1 + 3X2 ≤ 27
2X1 + 4X2 ≤ 32 Y1, Y2 ≥ 0
Solve the dual. What is the objective
X1 , X 2 ≥ 0 function value of the dual at the optimum?
Write the dual of the problem

Maximize 6X1 + 8X2


Subject to
X1 + X2  10
2X1 + 3X2  25
X1 + 5X2  35
X1, X2  0
Maximize 500X1 + 600X2
Primal solution is (2, 7) Z = 5200
3X1 + 3X2 ≤ 27
2X1 + 4X2 ≤ 32 Dual solution is (400/3, 50) W = 5200
X1 , X 2 ≥ 0
When 27 becomes 28, the solution is (8/3,
Minimize 27Y1 + 32Y2 20/3) Z = 5333.33
3Y1 + 2Y2 ≥ 500 Increase in Z is 1333.33 = 400/3 = Y1 at
3Y1 + 4Y2 ≥ 600 optimum
Y1, Y2 ≥ 0
When 32 becomes 33, the solution is (3/2,
15/2) with Z = 5250. The increase is 50
What happens if the RHS
which is the value of Y1 at optimum
of the first constraint
becomes 28?
Maximize 6X1 + 8X2 Primal solution is (5, 5) Z = 70
Subject to Dual solution is (2, 2, 0) W = 70
X1 + X2  10
2X1 + 3X2  25 When 35 becomes 36, the solution is (5, 5) Z =
X1 + 5X2  35 70. Increase in Z is 0 = Y3 at optimum
X1 , X 2  0
Is it because the third resource was not fully
Minimize 10Y1 + 25Y2 + 35Y3 utilized?
Y1 + 2Y2 + Y3 ≥ 6
Y1 + 3Y2 + 5Y3 ≥ 8 Dual has two constraints and only two
variables will be in the solution. Is this true?
Y1, Y2, Y3 ≥ 0 Then the primal has three constraints. There
are only 2 variables. What happens?
Maximize 500X1 + 600X2
Can we read the dual solution from excel
3X1 + 3X2 ≤ 27 output of the primal? (without actually
2X1 + 4X2 ≤ 32 solving the dual)
X1 , X 2 ≥ 0
Concept of shadow price
Minimize 27Y1 + 32Y2
3Y1 + 2Y2 ≥ 500 Price of an extra unit of resource at
3Y1 + 4Y2 ≥ 600 optimum
Y1, Y2 ≥ 0
Marginal worth of a unit at optimum

Can we see the dual in the algebraic form


of simplex?
Maximize 500X1 + 600X2 Can we see the dual in the algebraic form
3X1 + 3X2 ≤ 27 of simplex?
2X1 + 4X2 ≤ 32
Primal solution is (2, 7) Z = 5200
X1 , X 2 ≥ 0
Dual solution is (400/3, 50) W = 5200
Minimize 27Y1 + 32Y2
3Y1 + 2Y2 ≥ 500   400
3Y1 + 4Y2 ≥ 600 𝑍 =5200 − 𝑋 − 5 0 𝑋4
3 3
Y1, Y2 ≥ 0
Relationship between Y1 and X3 as well as
between Y2 and X4
Economic Interpretation of the dual

Use both the examples

Explain Y3 = 0 using the second example

Marginal worth vs notional worth

New product idea and worth

Case: Auto Assembly


Marginal worth vs notional worth

Maximize 500X1 + 600X2


3X1 + 3X2 ≤ 27
2X1 + 4X2 ≤ 32
X1 , X 2 ≥ 0

Worth of resources of product 1 = 400/3 x 2 + 40 x 2 = 500. I sell for 500.


Where is my profit?

You can consider your organization as a resource and hence the profit
comes. I buy these resources at the notional cost and want to minimize the
total cost of the resources. The constraint becomes 3Y1 + 2Y2 ≥ 500;
otherwise I wont produce. Finally it becomes equal.
 Sensitivity table in Excel output

 Dual in game theory formulation and capital investment example

Strategy for Stone paper Scissor


Player A
Stone 0 -2 3
paper 3 0 -2
Scissor -2 4 0
Let p1, p2, p3 be the proportion of times A plays the three strategies. If B consistently
plays stone, A’s profit will be 3p2 - 2p3. If B consistently plays paper, A’s profit will be -2p1
+ 4p3. If B consistently plays scissor, A’s profit will be 3p1 - 2p3. Since A and B are
“rational”, A will play in such a way to Maximize A’s minimum profit (because B would
allow minimum profit). A plays what is called “maximin’ strategy.
Maximize u
u ≤ 3p2 - 2p3;
u ≤ -2p1 + 4p3;
u ≤ 3p1 - 2p2;

p1 + p2 + p3 = 1.
p1, p2, p3 ≥ 0. u unrestricted in sign
Let q1, q2, q3 be the proportion of times B plays the three strategies. If A consistently
plays stone, B’s loss will be -2q2 + 3q3. If A consistently plays paper, B’s loss will be 3q1 –
2q2. If A consistently plays scissor, B’s loss will be -2q1 + 4q2. Since A and B are “rational”,
B will play in such a way to Minimize B’s maximum loss (because A would like to give
maximum loss to B). B plays what is called “minimax’ strategy.
Minimize v
v ≥ -2q2 + 3q3;
v ≥ 3q1 – 2q3;
v ≥ -2q1 + 4q2;

q1 + q2 + q3 = 1.
q1, q2, q3 ≥ 0. v unrestricted in sign
capital Investment example

Maximize 44X1 + 70X2 + 35X3 + 35X4 + 88X5 + 35X6 + l3 – b3


12X1 + 54X2 + 6X3 + 6X4 +30X5 + 6X6 + l1 – b1 ≤ 35
3X1 + 10X2 + 6X3 + 2X4 +35X5 + 10X6 – 1.05l1 +1.05b1 + l2 – b2 ≤ 20
5X1 + 4X2 + 6X3 + 5X4 +10X5 + 4X6 – 1.05l2 +1.05b2 + l3 – b3 ≤ 20
0 ≤ Xj ≤ 1

Solution: X1 = X3 = X4 = X5 = X6 = 1; b1 = 25, b2 = 62.25, b3 = 75.36


and Z = 161.63

Dual solution is Y1 = 1.102, Y2 = 1.05, Y3 = 1, u1 = 22.62, u3 = 16.08,


u4 = 21.28, u5 = 8.17 u6 = 13.88 with W = 161.63

You might also like