Professional Documents
Culture Documents
Carla Gomes
Module 7b
Duality and Sensitivity Analysis
Economic Interpretation of Duality
(slides adapted from: M. Hillier’s, J. Orlin’s, and H. Sarper’s)
Post-optimality Analysis
• Post-optimality – very important phase of
modeling.
• Duality plays and important role in post-
optimality analysis
• Simplex provides several tools to perform
post-optimality analysis
Post-optimality analysis for LP
Task Purpose Technique
xi >= 0 , (i =1,2,…,n)
• Resources – m (plants)
• Activities – n (2 products)
• Wyndor Glass problem optimal product mix ---
allocation of resources to activities i.e., choose the levels
of the activities that achieve best overall measure of
performance
3 D + 2 W = 18
8
(5/3,13/2)
D=4
2w=13 Z=3(5/3)+5(13/2)=37.5
6 2 W =12 Z=3(2)+5(6)=36
∆ Z=1.5
= y2 *
(2,6)
4
Feasible
Why is y1*=0?
region
2
0 2 4 6 8 D
Production rate for doors
Economic Interpretation of Dual Variables
(5/3,13/2) D=4
2w=13 Z=3(5/3)+5(13/2)=37.5
6 2 W =12 Z=3(2)+5(6)=36
∆ Z=1.5
= y2 *
(2,6)
4
0 2 4 6 8 D
Production rate for doors
3 D + 2 W = 18
8
D=4
(5/3,13/2) 2w=13 Z=3(5/3)+5(13/2)=37.5
6 2 W =12 Z=3(2)+5(6)=36
∆ Z=1.5
= y2 *
(2,6) If b2 < 6 the solution will no longer
4
vary proportionally. The optimal
Feasible b2=6 solution varies proportionally to the
region variation in b2 only if 6 <= b2 <=18.
2
In other words, the current basis remains
optimal for 6 ≤ b2 ≤ 18, but the decision
variable values and z-value will change.
0 2 4 6 8 D
Production rate for doors
• A dual variable yi* gives us the rate at which Z could be
increased by increasing the amount of resource i slightly.
• However this is only true for a small increase in the
amount of the resource. I.e., this definition applies only if
the change in the RHS of constraint i leaves the current
basis optimal. It also assumes everything else stays the
same.
• Another interpretation of yi* is: if a premium price must
be paid for the resource i in the market place, yi* is the
maximum premium (excess over the regular price) that
would be worth paying.
Optimal Basis in the Wyndor Glass
Problem
• How can we characterize (verbally) the
optimal basis of the Wyndor Glass problem?
Optimal solution
Feasible
4
region
P = 1500 = 300D + 500W
isoprofit line
0 2 4 6 8 10 D
Production rate for doors
How much can c1 vary until the slope of the iso-profit line
equals the slope of constraint 2 and constraint 3?
• How much can c1
vary until the slope of c1 D 0 c1 0
the iso-profit line 5
equals the slope of
constraint 2 and
c1 D 3 c115
5 2 2
constraint 3? 0 c1 7.5
• Slope of constraint
2 0
• Slope of constraint 3
-3/2
Importance of Sensitivity Analysis
z = 5 x1 + 4.5x2
x1 = 6 3/7 + 2/7
6 x1 + 5 x2 = 60 +
10 x1 + 20 x2 = 150 x2 = 4 2/7 – /7
z = 51 3/7 + 11/14
How do we interpret the intervals?
• If we change one coefficient in the RHS, say
production capacity, by the “basis” remains
optimal, that is, the same equations remain
binding.
• So long as the basis remains optimal, the shadow
prices are unchanged.
• The basic feasible solution varies linearly with.
If is big enough or small enough the basis will
change.
Illustration with the glass example:
max 5 x1 + 4.5 x2 + 6 x3 ($100s)
s.t 6 x1 + 5 x2 + 8 x3 60 (prod. cap. in hrs)
10 x1 + 20 x2 + 10 x3 150 (wareh. cap. in ft2)
x1 8 (6-0z. glass dem.)
x1 0, x2 0, x3 0
4.5
Reduced cost of x2 = - 5 x 11/14
- 20 x 1/35
- 0 x0
= 4.5 – 55/14 – 4/7 = 0
Pricing Out of x3
shadow price
max 5 x1 + 4.5 x2 + 6 x3 ($100s)
s.t 6 x1 + 5 x2 + 8 x3 60 ……11/14
10 x1 + 20 x2 + 10 x3 150
……1/35
1 x1 8
x1 0, x2 0, x3 0 …….0
6
Reduced cost of x3 = - 8 x 11/14
- 10 x 1/35
- 0 x0
= 6 – 44/7 – 2/7 = -4/7
Can we use pricing out to figure out
whether a new type of glass should be
produced? shadow price
max 5 x1 + 4.5 x2 + 7 x4 ($100s)
s.t 6 x1 + 5 x2 + 8 x4 60 ……11/14
10 x1 + 20 x2 + 20 x4 150 ……1/35
1 x1 8
…….0
x1 0, x2 0, x4 0
7
Reduced cost of x4 = - 8 x 11/14
- 20 x 1/35
- 0 x0
= 7 – 44/7 – 4/7 = 1/7
Pricing Out of xj
shadow price
max 5 x1 + 4.5 x2 + cj xj ($100s)
s.t 6 x1 + 5 x2 + a1j xj 60
……y1
10 x1 + 20 x2 + a2j xj 150
……….. ……y2
………. + amjxj = bm
………
x1 0, x2 0, x3 0
……ym
Reduced cost of xj = ?
Brief summary on reduced costs
• The reduced cost of a non-basic variable xj is the
“increase” in the objective value of requiring that
xj >= 1.
• The reduced cost of a basic variable is 0.
• The reduced cost can be computed by treating
shadow prices as real prices. This operation is
known as “pricing out.”
• Pricing out can determine if a new variable would
be of value (and would enter the basis).
Summary
• The shadow price is the unit change in the optimal
objective value per unit change in the RHS.
• The shadow price for a “ 0” constraint is called the
reduced cost.
• Shadow prices usually but not always have economic
interpretations that are managerially useful.
• Non-binding constraints have a shadow price of 0.
• The sign of a shadow price can often be determined by
using the economic interpretation
• Shadow prices are valid in an interval.
• Reduced costs can be determined by pricing out
Reduced Costs
• The reduced cost of a variable x is the shadow
price of the “x 0” constraint. It is also the
negative of cost coefficient for x in the final
tableau.
c1 cost
Reduced
Dual or Shadow is the amount the
prices are the objective function
amount the coefficient for
optimal z-value variable i would
improves if the have to be
rhs of a Dual variables increased for
constraint is there to be an
increased by one b2
alternative
unit (assuming optimal solution.
no change in
basis). More later…
MPL – Sensitivity analysis info
Allowable ranges
c1 (w/o changing basis)
for the x1 coefficient
(c1) is:
0 c1 7.5
Allowable ranges – in
terms of increase and
decrease
(w/o changing basis)
for the x1 coefficient
(c1) is:
0 c1 7.5
The Computer and Sensitivity Analysis
• Consider the following maximization problem. Winco sells
four types of products. The resources needed to produce one
unit of each are:
Product Product Product Produc Availabl
1 2 3 t4 e
Raw 2 3 4 7 4600
material
Hours of 3 4 5 6 5000
labor
Sales price $4 $6 $7 $8
To meet customer demand, exactly 950 total units must be
produced. Customers demand that at least 400 units of product 4
be produced. Formulate an LP to maximize profit.
Let xi = number of units of product i produced by Winco.
• The Winco LP formulation:
NO. ITERATIONS= 4
RANGES IN WHICH THE BASIS IS UNCHANGED:
LINDO sensitivity
OBJ COEFFICIENT RANGES
analysis example(s).
VARIABLE CURRENT ALLOWABLE ALLOWABLE
COEF INCREASE DECREASE
Allowable range (w/o
X1 4.000000 1.000000 INFINITY
changing basis) for
X2 6.000000 0.666667 0.500000
the x2 coefficient
X3 7.000000 1.000000 0.500000
(c2) is:
X4 8.000000 2.000000 INFINITY
5.50 c2 6.667
RIGHTHAND SIDE RANGES
Allowable range (w/o ROW CURRENT ALLOWABLE ALLOWABLE
changing basis) for RHS INCREASE DECREASE
the rhs (b1) of the first 2 950.000000 50.000000 100.000000
constraint is: 3 400.000000 37.500000 125.000000
4 4600.000000 250.000000 150.000000
850 b1 1000
5 5000.000000 INFINITY 250.000000
MAX 4 X1 + 6 X2 + 7 X3 + 8 X4
Shadow prices SUBJECT TO
2) X1 + X2 + X3 + X4 = 950
are shown in the 3) X4 >= 400
Dual Prices 4) 2 X1 + 3 X2 + 4 X3 + 7 X4 <= 4600
5) 3 X1 + 4 X2 + 5 X3 + 6 X4 <= 5000
section of END
LINDO output. LP OPTIMUM FOUND AT STEP 4
Assuming the allowable range of the rhs is not violated, shadow (Dual) prices
show: $3 for constraint 1 implies that each one-unit increase in total demand
will increase net sales by $3. The -$2 for constraint 2 implies that each unit
increase in the requirement for product 4 will decrease revenue by $2. The $1
shadow price for constraint 3 implies an additional unit of raw material (at no
cost) increases total revenue by $1. Finally, constraint 4 implies any additional
labor (at no cost) will not improve total revenue.
Shadow price signs
1. Constraints with symbols will always have
nonpositive shadow prices.
2. Constraints with will always have nonnegative
shadow prices.
3. Equality constraints may have a positive, a
negative, or a zero shadow price.
Managerial Use of Shadow Prices
The managerial MAX 4 X1 + 6 X2 + 7 X3 + 8 X4
SUBJECT TO
significance of shadow 2) X1 + X2 + X3 + X4 = 950 raw
prices is that they can 3) X4 >= 400
material
4) 2 X1 + 3 X2 + 4 X3 + 7 X4 <= 4600
often be used to 5) 3 X1 + 4 X2 + 5 X3 + 6 X4 <= 5000
determine the END
labor
maximum amount a LP OPTIMUM FOUND AT STEP 4
manager should be OBJECTIVE FUNCTION VALUE
willing to pay for an 1) 6650.000
additional unit of a VARIABLE VALUE REDUCED COST
resource. Reconsider X1 0.000000 1.000000
X2 400.000000 0.000000
the Winco to the right. X3 150.000000 0.000000
X4 400.000000 0.000000
What is the most ROW SLACK OR SURPLUS DUAL PRICES
Winco should be 2) 0.000000 3.000000
3) 0.000000 -2.000000
willing to pay for 4) 0.000000 1.000000
additional units of raw 5) 250.000000 0.000000