Week

Sensitivity
9 Analysis
Opera)ons Research I
 A Graphical Introduction to Sensitivity Analysis

•  Sensitivity analysis concerns with how changes in an LP’s parameters

affect the optimal solution

Resource 1 (9inishing) constraint

Resource 2 (carpentry) constraint
Resource 3 (demand) constraint

Where,
x1 = number of product 1 produced per week
x2 = number of product 2 produced per week

The optimal solution to this problem is z = 180, x1 = 20, x2 = 60, and it has x1, x2, and
s3 (the slack variable for the demand constraint) as basic variables.
How would changes in the problem’s objective function coefKicients or right-hand
sides change this optimal solution?
Graphical Analysis – Objective Function CoefKicients

In this example, if the change

in the objective function
coefKicients causes the
isoproKit line to be Klatter
(larger gradient) than
carpentry constraint, the
optimal solution will change.

Because the slope of the
carpentry constraint is -1, the
isoprofit lines will be flatter
than the carpentry constraint
if –c1/2 > -1 or c1<2, and the
current basis will no longer
be optimal.
Graphical Analysis – Change on RHS

If we change the value of b1,

then as long as the point
where the finishing and
carpentry constraints are
binding remains feasible, the
optimal solution will still occur
where the finishing and
carpentry constraints
intersect.
 Shadow Price

•  It is often important for managers to determine how a change in a

constraint’s right-hand side changes the LP’s optimal z-value.
•  With this in mind, we define the shadow price for the ith constraint of
an LP to be the amount by which the optimal z-value is improved
(increased in a max problem and decreased in a min problem) if the
right-hand side of the ith constraint is increased by 1.
•  This definition applies only if the change in the right-hand side of
constraint i leaves the current basis optimal.
 Simplex Interpretation

Let’s take the following as an example:

Max x1 + 3x2
ST  2x1 + 3x2 ≤ 8
− x1 + x2 ≤ 1
     x1, x2 , ≥ 0

6
7
 Shadow Price
Max x1 + 3x 2
ST 2x1 + 3x 2 + x 3 =8
− x1 + x 2 + x4 = 1
x1 , x 2 , x 3, x 4 ≥ 0 b1 is increased by 1 unit to 9

11/5

Shadow price for resource Z increased by 4/5 (shadow

1 b1 per unit price) à 7 + 4/5 = 39/5
 An Example (More)
Iteration 1:
z x1 x2 x3 x4 x5 x6 RHS
z 1 1 1 -4 0 0 0 0
x ...4 0 1 1 2 1 0 0 9
x ...5 0 1 1 -1 0 1 0 2
x ...6 0 -1 1 1 0 0 1 4
Iteration 2:
z x1 x2 x3 x4 x5 x6 RHS
z 1 -3 5 0 0 0 4 16
x ...4 0 3 -1 0 1 0 -2 1
x ...5 0 0 2 0 0 1 1 6
xx ...3 0 -1 1 1 0 0 1 4
Iteration 3:
z x1 x2 x3 x4 x5 x6 RHS
z 1 0 2 0 1 0 2 17
x ...1 0 1 -1/3 0 1/3 0 -2/3 1/3
x ...5 0 0 2 0 0 1 1 6
x ...3 0 0 2/3 1 1/3 0 1/3 13/3
9
 Simplex Tableau Interpretation (1)
•  In each iteration we multiply the tableau with B-1. Thus, I
in the initial tableau become B-1 in the next iteration.
Example: Iteration 2 (table 3) in the above example
⎡... ... ...⎤ ⎡ 13 0 − 2 3 ⎤
B = ⎢⎢... ... ...⎥⎥ ⇒ B −1 = ⎢⎢ 0 1 1 ⎥⎥
⎢⎣... ... ...⎥⎦ ⎢⎣ 13 0 13 ⎥⎦

z = c B B−1 b − ∑ ( z j − c j ) x j
j ∈R

b is the available resource

∂z ∂z
€ = cBB = w ⇒
−1
= wi
∂b ∂bi
 Simplex Tableau Interpretation (2)
∂z
= change in Z when resource i is increased by 1
∂bi
unit.
= shadow price of resource i

w = c B B −1 ⇒ elements of row 0 in the current

tableau under the variables with identity matrix in
the initial tableau.
Example for the third iteration in the example:
z x1 x2 x3 x4 x5 x6 RHS
z 1 0 2 0 1 0 2 17
x1 0 1 -1/3 0 1/3 0 -2/3 1/3
x5 0 0 2 0 0 1 1 6
x3 0 0 2/3 1 1/3 0 1/3 13/3
∂z ∂z
w= = ...... ; = w1 = ......
∂b ∂b1
 Simplex Tableau Interpretation (3)
∂z
= −(z j − c j ) For xj non-basic
∂x j
= the change in z when xj changes

'
•  x B = B−1 b − B−1N x N = B−1 b − ∑ a j x j
j ∈R

∂xB −1 ∂xB
=B ; = The jth column in B-1
∂b ∂b j
∂xBi
= The ith row of B-1
∂b
 Iterpretasi dari Tableau Simpleks (4)
from the example, define:
z x1 x2 x3 x4 x5 x6 RHS
z 1 0 2 0 1 0 2 17
x1 0 1 -1/3 0 1/3 0 -2/3 1/3
x5 0 0 2 0 0 1 1 6
x3 0 0 2/3 1 1/3 0 1/3 13/3

∂z ∂xB ∂x5 ∂x5

= ...... ; = ...... ; = ...... ; = ......
∂x2 ∂b2 ∂b ∂b2
• 
∂xB
= −aʹj : the column below xj
∂x j

Based on the 3rd iteration, determine:

∂xB ∂ x3
= ...... ; = ......
∂x2 ∂x3
More Examples on Table Interpretation
 Implementation: An Example
Max 5 x1 + 3 x2
ST ?
In a certain iteration, the following tableau is resulted:
z x1 x2 x3 x4 RHS
z 1 b 1 f g 10
x3 0 c 0 1 1/5 2
x1 0 d e 0 1 a
x3, x4 = slack

•  Determine the value of a – g

–  x1 = basic / nonbasic ?
–  x2 = basic / non basic ?
–  x3 = basic / non basic ?
 z x1 x2 x3 x4 RHS
z 1 b 1 f g 10
x3 0 c 0 1 1/5 2
x1 0 d e 0 1 a

–  b = …… –  c = ……

–  d = …… –  e = ……

–  f = ……
–  a = ……
–  z = …… ⇒ '
–  z2 = …… = cB a 2
–  z2 – c2 = …… ⇒
–  e = ……
–  g = ……
 z x1 x2 x3 x4 RHS
z 1 0 1 0 5 10
x3 0 0 0 1 1/5 2
x1 0 1 4/5 0 1 2

–  Is the tableau optimal?

w1 = ; w2 =

–  If you have additional money, which resource would you

buy?

–  At what price would you buy it?

–  Would you sell resource 1 @ 2$/unit? How many would

you sell?
 More Example
Max 2x1 +12x2 + 7x3                   ( profit)
ST x1 + 3x2 + 2x3 ≤ 10000
2x1 + 2x2 + x3 ≤ 4000
x1 , x 2 , x3 ≥ 0
The optimal tableau is as follows:
z x1 x2 x3 x4 x5 RHS
z 1 12 2 0 0 7 28000
x4 0 -3 -1 0 1 -2 2000
x3 0 2 2 1 0 1 4000

§  A new product x6 is considered to be produced with a profit

of $6 per unit. To produce it, 2 units of R1 and 2 units of
R2 will be required. Should we produce it? Why?
 Analysis

z x1 x2 x3 x4 x5 RHS
z 1 12 2 0 0 7 28000
x4 0 -3 -1 0 1 -2 2000
x3 0 2 2 1 0 1 4000

§  What is the minimum profit so that this product can be

produced?

§  Let say you have a capacity to buy Δ unit of resource 2 at

5$ per unit. Would you buy it?

§  How many is Δ so that the current tableau remains

optimal?
Excel Solution
Week 10 – Sensitivity Analysis

Analysis using Software Output

 Winco Case

Winco sells four types of products. The resources needed to produce one unit of
each and the sales prices are given in Table 2. Currently, 4,600 units of raw
material and 5,000 labor hours are available. To meet customer demands, exactly
950 total units must be produced. Customers also demand that at least 400 units
of product 4 be produced. Formulate an LP that can be used to maximize Winco’s
sales revenue.







Let xi = number of units of
product i produced by Winco.
•  The solu8on for the
problem was
generated from
Lindo as shown on
the le< side.

•  Now using Solver

(Ms. Excel), find the
solu8on and
sensi8vity report
for the same
problem.
 Interpretation of Objective Function CoefKicients
Sensitivity Analysis

a)  Suppose Winco raises the price of product 2 by 50¢ per unit. What
is the new optimal solution to the LP?
b)  Suppose the sales price of product 1 is increased by 60¢ per unit.
What is the new optimal solution to the LP?
c)  Suppose the sales price of product 3 is decreased by 60¢. What is
the new optimal solution to the LP?
 Sensitivity Analysis for a Maximization Example

•  Hartono Electronics considers producing four potential products: VCRs,

stereos, televisions, and DVD players. They all require electronic components
(cost $7/unit), non electronic components (cost $5/unit) and assembly time
(cost $10/hour).

VCR STEREO TV DVD SUPPLY

Electronic components 3 4 4 3 4700

Non electronic components 2 2 4 3 4500

Assembly 8me 1 1 3 2 2500

Selling price $70 $80 $150 $110

•  The pro9it per unit then follows: VCR = $29, stereo $32, TV $72, DVD $54
(calculation example VCR proKit= 70-3(7)-2(5)-1(10)=29)
Problem Formulation
Excel-Solver Report
 What-If Analysis

1.  What is the impact on pro9it of a change in the supply of non-electronic

components?
2.  What is the impact on pro9it if we could increase the supply of electronic
components by 400 units?
3.  What would happen if we could increase the supply of electronic components
by 4000 units?
4.  What would happen if the supplier of these 400 units (no. 2) wants $8 per
unit (rather than the current cost of $7 per unit)?
5.  Assume we have an opportunity to obtain 250 additional hours of assembly
time @a price of $15 per hour. Should we take it?
6.  If we force the production of VCRs, what would be the impact on total pro9it?
Alternatively, how pro9itable must VCRs become before Hartono should
consider producing them?
7.  Assume there is some uncertainty in the price for DVD players. For what
range of prices will the current production plan be optimal? If DVD players
sold for $106, what would be Hartono’s new total pro9it?
 Simultaneous Changes (100% rule)

•  It is possible to analyze the impact of simultaneous changes on optimal

solution under speci9ic condition:
Σ (change/Allowable change) ≤ 1

•  Note: NOT applicable if you have simultaneous changes in OFC and RHS
 Pricing Out

•  Let say Hartono wants to produce Home Theatre that requires 5 units
of electronic components, 4 units of non electronic components and 4
units of assembly time with selling price $175 per unit.
Should Hartono produce it or not? Why?