Professional Documents
Culture Documents
AY 2021 Sem 1 E210 - W03 - IS - SLIDES
AY 2021 Sem 1 E210 - W03 - IS - SLIDES
Lesson 03
Product Mix and Distribution Planning
2
OFFICIAL (CLOSED) \ NON-SENSITIVE
IS Lecture (1)
Case Scenario
OFFICIAL (CLOSED) \ NON-SENSITIVE
• How should PrettyFit plan the product mix to maximize profit with the limited resources?
• How would the available resources and the unit profit of each product affect the overall
profit from producing the products?
• How should PrettyFit develop a distribution plan to minimize the distribution cost while
considering both available supplies and customer demands?
4
OFFICIAL (CLOSED) \ NON-SENSITIVE
5
OFFICIAL (CLOSED) \ NON-SENSITIVE
Product Machine Time (Hour) Cloth (m2) Labour Time (Hour) Unit Profit ($)
Midi-dress 0.3 1.2 1.5 24
Mini-dress 0.6 0.8 2 36
Maxi-dress 0.9 1.5 1.8 42
• Kate recalls that the new product mix can be determined using a mathematical model
and Excel Solver.
6
OFFICIAL (CLOSED) \ NON-SENSITIVE
7
OFFICIAL (CLOSED) \ NON-SENSITIVE
8
OFFICIAL (CLOSED) \ NON-SENSITIVE
9
OFFICIAL (CLOSED) \ NON-SENSITIVE
IS Lecture (2)
Mathematical Models
• Set of mathematical relationships, equations and
logical assumptions as a representation of real world
decision making problem.
11
OFFICIAL (CLOSED) \ NON-SENSITIVE
Mathematical Models
• 3 stages in the application of mathematical
models to real-world problems:
Translating problem situation into a formal
mathematical model,
Solving mathematical problem, and
Interpreting solution for the original situation
12
OFFICIAL (CLOSED) \ NON-SENSITIVE
13
OFFICIAL (CLOSED) \ NON-SENSITIVE
15
OFFICIAL (CLOSED) \ NON-SENSITIVE
LP Model Formulation
• Helpful questions to identify the major components of
the Linear Programming model:
1. What am I trying to decide? (Decision Variables)
E.g. Number of Product A to make, Number of staff to hire
16
OFFICIAL (CLOSED) \ NON-SENSITIVE
LP Formulation - Decision Variables & Objective
Function
LP Formulation - Constraints
• Each constraint in an LP represents a relation between two
measurable quantities: Limited resource (≤), Minimum
performance (≥), and Conservation/ balance (=).
• Example: A total of 150 labour hours are available for the
production of X1 and X2
20X1 + 10X2 ≤ 150
19
OFFICIAL (CLOSED) \ NON-SENSITIVE
S=20T
S>=15T
F/(F+M)<=50% or F<=M
20
OFFICIAL (CLOSED) \ NON-SENSITIVE
• In the coming week, there are a maximum of 160 machine hours and 240 labour
hours available, respectively.
• Danny recalls that the product mix can be determined using a mathematical
model and a graphical approach.
21
OFFICIAL (CLOSED) \ NON-SENSITIVE
22
OFFICIAL (CLOSED) \ NON-SENSITIVE
24
OFFICIAL (CLOSED) \ NON-SENSITIVE
Decision Variables
Objective Function
Maximize profit Z = 120X1 + 150X2
Constraints
IS Lecture (3)
Graphical Method
OFFICIAL (CLOSED) \ NON-SENSITIVE
Then either:
3. Plot the objective function and move this line in parallel away from the
origin. Locate the extreme point of the feasible region.
Or:
2. Substitute these values into the objective function to find the set of values
that results in maximum/minimum Z value.
Note: It may be possible to read the solution values directly from the graph
instead of solving as simultaneous equations.
(Watch the video https://docs.google.com/file/d/0Bz-Yttbj61bxeW16a3R1NEQ5dVE/edit?usp=sharing)
27
OFFICIAL (CLOSED) \ NON-SENSITIVE
a) Match the lines in the graphs with the constraints that you have
formulated.
28
OFFICIAL (CLOSED) \ NON-SENSITIVE
a) Match the lines in the graphs with the constraints that you have
formulated.
• To understand the inequality, substitute values such that the LHS is not
equal to the RHS and then see where the point falls in the graph.
<= : below the line; towards the left side of the line
>= : above the line; towards the right side of the line
29
OFFICIAL (CLOSED) \ NON-SENSITIVE
(Type A)
30
OFFICIAL (CLOSED) \ NON-SENSITIVE
100
80 Infeasible region
Machine Hours Constraint:
0.8X1 + 1.6X2 ≤ 160
60
40
20
Feasible region X2 ≥ 0
0
0 50 100 150 200 250
X1(Type A)
31
OFFICIAL (CLOSED) \ NON-SENSITIVE
40 Feasible
region
20
X2 ≥ 0
0
0 50 100 150 200 250
X1 (Type A)
32
OFFICIAL (CLOSED) \ NON-SENSITIVE
a) What happens to Z when you move this line in parallel away from the origin?
b) Where is the furthest you can move without leaving the feasible solution region?
c) What is the significance of this extreme (furthest) point in the feasible region?
33
OFFICIAL (CLOSED) \ NON-SENSITIVE
a) What happens to Z when you move this line in parallel away from the origin?
b) Where is the furthest you can move without leaving the feasible solution region?
c) What is the significance of this extreme (furthest) point in the feasible region?
a) Z (the profit) changes when the objective function line moves. And since the
objective is to maximize profit, the line should be moved as far right or upwards
as possible without leaving the feasible region.
b) As the objective function line moves further, its value (right-hand side of equation)
increases. Optimal solution is found by moving the objective function line right
and upwards (keeping to its gradient) until the extreme point is reached in the
feasible solution region.
c) This is the optimal solution that will give the maximum objective function value.
34
OFFICIAL (CLOSED) \ NON-SENSITIVE
80
Machine Hours Constraint:
0.8X1+ 1.6X2 ≤ 160
60
40
20
X2 ≥ 0
0
0 50 100 150 200 250
Lines parallel to the objective X1 (Type A)
function: Z = 120X1 +150X2
35
OFFICIAL (CLOSED) \ NON-SENSITIVE
100
Extreme point: X1 = 20; X2= 90
Machine hours used = 160 hours
80 Labour hours used = 240 hours
Profit = 120x20 + 150x90 = $15,900
60
20
X2 ≥ 0
0
0 50 100 150 200 250
X1 (Type A)
36
OFFICIAL (CLOSED) \ NON-SENSITIVE
20
X2 ≥ 0
0
0 50 100 150 200 250
X1 (Type A)
37
OFFICIAL (CLOSED) \ NON-SENSITIVE
Solutions
• The optimal solution at which profit is maximized
is X1= 20, X2= 90.
• In other words, we recommend to make 20 units
of Type A and 90 units of Type B per week which
will yield a maximum profit of $15,900.
Profit = 120x20 + 150x90 = $15,900
IS Lecture (4)
Sensitivity Analysis
- Shadow Price and Range of Feasibility
OFFICIAL (CLOSED) \ NON-SENSITIVE
40
OFFICIAL (CLOSED) \ NON-SENSITIVE
41
OFFICIAL (CLOSED) \ NON-SENSITIVE
Sensitivity Analysis
• Sensitivity analysis aids in answering questions such as:
Range of optimality - The range of values that the objective
function coefficients can assume without changing the optimal
solution.
Range of feasibility - The range of values for the right-hand side
of a constraint, in which the shadow price for the constraint remains
unchanged - the valid range of shadow price.
Reduced Cost - The impact of forcing a variable which is
currently zero to be non-zero on the optimal objective function
value.
Shadow price - The impact of increasing or decreasing the right-
hand side value of various constraints on the optimal objective
function value.
The impact of changing the constraint coefficients on the optimal
solution to the problem.
42
OFFICIAL (CLOSED) \ NON-SENSITIVE
Sensitivity Analysis
• The information and insights obtained through
sensitivity analysis are valuable to management
They provide an indication of the degree of flexibility
that is inherent in an operating environment.
• Almost all commercial software for linear
programming provides
Right-hand side ranging information
Objective coefficient ranging information.
Other sensitivity analysis, such as adding new variable
and constraint, is dealt by solving the modified linear
programming problem.
43
OFFICIAL (CLOSED) \ NON-SENSITIVE
44
OFFICIAL (CLOSED) \ NON-SENSITIVE
45
OFFICIAL (CLOSED) \ NON-SENSITIVE
20
X2 ≥ 0
0
0 50 100 150 200 250
X1 (Type A)
46
OFFICIAL (CLOSED) \ NON-SENSITIVE
47
OFFICIAL (CLOSED) \ NON-SENSITIVE
20
X2 ≥ 0
0
0 50 100 150 200 250
49
OFFICIAL (CLOSED) \ NON-SENSITIVE
140 X1 ≥ 0
Max Z = 120X1 +150X2 = $15,900
120 Beyond this point, the Labour
Hours constraint becomes
Machine Hours Constraint: non-binding.
X2 (Type B)
100
0.8X1+ 1.6X2 ≤ 160 At this point the profit is
$24,000
80
Change in profit = [Shadow
Price]*[Change in RHS value]
60
= 22.5*(600-240) = $8,100
40
New Optimal
20 Solution: (200,0)
X2 ≥ 0
0
0 50 100 150 200 250
50
OFFICIAL (CLOSED) \ NON-SENSITIVE
51
OFFICIAL (CLOSED) \ NON-SENSITIVE
140 X1 ≥ 0
Below this 200 labour hours, Machine Hours
Max Z = 120X1 + 150X2 = $15,000 constraint becomes non-binding. Only Labour Hours
120 constraint is binding.
New Optimal
Solution: (0,100) At this point the profit is $15,000
X2 (Type B)
100
Change in profit = [Shadow Price]*[Change in RHS
value]
80
= 22.5*(200-240) = -$900
60
20
X2 ≥ 0
0
0 50 100 150 200 250
When labour hours are reduced,
the right-hand side of the labour
X1 (Type A)
hours constraint decreases.
52
OFFICIAL (CLOSED) \ NON-SENSITIVE
53
OFFICIAL (CLOSED) \ NON-SENSITIVE
100 In this case, the labour hours should be within 200 to 600 hours for
the shadow price of the labour hours constraint to be $22.5.
80
60
40
20
X2 ≥ 0
0
0 50 100 150 200 250
X1 (Type A)
54
OFFICIAL (CLOSED) \ NON-SENSITIVE
IS Lecture (5)
Sensitivity Analysis
- Range of Optimality
OFFICIAL (CLOSED) \ NON-SENSITIVE
56
OFFICIAL (CLOSED) \ NON-SENSITIVE
60
40
20
X2 ≥ 0
0
0 50 100 150 200 250
If the unit profit for Type A increases to $225, the X1 (Type A)
objective function will be parallel to the Labour
What happens if the unit profit for
Hours constraint (multiple optimal solution).
Type A further reduces / increases
Max Z = 225X1 + 150X2
beyond this range?
Profit = $18,000
57
OFFICIAL (CLOSED) \ NON-SENSITIVE
60
40
20
X2 ≥ 0
0
0 50 100 150 200 250
If the unit profit for Type A increases to $450
Extreme point: (X1 = 80; X2=0)
X1 (Type A)
Max Z = 450X1 + 150X2
Profit = $24,000
58
OFFICIAL (CLOSED) \ NON-SENSITIVE
Validity of Model
• Verify the model – check the formula and assumptions
• Check that the model represents reality by comparing
the result with the expected outcome
Use historical data to reconstruct the past and then
determine how well the model and the resulting solution
would have performed if they had been used.
59
OFFICIAL (CLOSED) \ NON-SENSITIVE
Learning Outcomes
At the end of the lesson, students should be able to:
• Identify problems that can be solved using linear programming.
• Formulate a Linear Programming (LP) model by establishing the decision variables, objective
function and constraints.
• Solve a Linear programming (LP) problem with two decision variables using graphical method.
• Perform sensitivity analysis on the objective function coefficients and right-hand side values of the
constraints using graphical method.
• Interpret feasibility and optimality of an LP solution.
• Solve an LP model using MS Excel Solver.
• Interpret shadow price to determine the impact of changing constraint right-hand side value on the
optimal objective value.
• Explain range of optimality: the range of values that the objective function coefficients can assume
without changing the optimal solution.
• Explain range of feasibility: the range of values for the right-hand side of a constraint, in which the
shadow price for the constraint remains unchanged.
• Calculate Reduced Cost to determine the impact of forcing a decision variable which has zero value
to be non-zero on the objective function value.
• Identify a transportation problem that deals with the distribution of goods from multiple supply
sources to multiple demand destinations.
• Apply North-West Corner Rule and Lowest Cost Method to determine a feasible solution for a
transportation problem.
• Formulate the transportation problem that involves minimizing the cost of shipping goods from a set
of origins to a set of destinations as a linear programming model.
• Apply Excel Solver to find an optimal shipment allocation for the transportation problem to minimize
the total transportation cost.
• Compare the heuristic solutions with the solution of linear programming method.
60
OFFICIAL (CLOSED) \ NON-SENSITIVE
Decision
Operations LP – Integer Making using
Planning Programming Decision Tree
Overview
LP -
Transportation Network
Model Diagram and
Decision
Making with Min. Spanning
Dependent Tree
Decisions
LP formulation LP – Binary
Process Strategy and EXCEL Integer
& Capacity Solver with Programming Decision
Requirements Sensitivity Making using
analysis Utility Dissimilarity
Function Index and MST
Linear
Programming Decision
(LP) Graphical Making under
Method uncertainty
and risk
We are
here !
61