ISOM 2700: Operations Management
Session 6. Resource Management I
Capacity Planning
Lijian Lu
Dept. of ISOM, HKUST
1
This class
• Capacity planning
• Decision tree method
• Value of perfect information
2
Capacity Planning
Transformation Process
Inputs Goods or Services
Resources
Labor, machines, Cash…
3
Capacity Planning
Capacity is the (maximum) amount of output that a system is
capable of achieving over a specific period of time.
• Three levels of capacity planning and examples
– Strategic: Plant and equipment investments, etc.
– Tactical: Hiring, layoffs, new tools, minor equipment purchases
– Operational: Detailed planning and control, job assignment, etc.
4
Measures of Capacity
Business Inputs Outputs
Auto manufacturing Labor and machine hours # of cars per shift
Steel mill Furnace size Tons of steel per day
Oil refinery Refinery size Gallons of fuel per day
Farming Number of acres, cows Bushels of grain per acre
per year, gallons of milk
per day
Restaurant Number of tables, seating No. of meals served per day
capacity
Theater Number of seats No. of tickets sold per
performance
Retail sales Square feet of floor space Revenue generated per day
5
Economies & Diseconomies of Scale
Average
unit cost
of output 100-unit
plant
200-unit
plant 400-unit
300-unit
plant
plant
Diseconomies of Scale start
Volume
6
Best Operating Level
Average
unit cost
of output
Under-utilization Over-utilization
Economies of scale Diseconomies of scale
Best Operating Level
Volume
7
Decision Making Process
• Specify objectives
Formulation
• Develop alternatives
• Analyze and compare alternatives
Analysis
• Select the best alternative
Implementation • Implement and monitor chosen
alternative
Evaluating Alternatives: Decision Tree Method
8
Glass Factory Example
A glass factory specializing in crystal is experiencing a
substantial backlog, and the firm's management is
considering three courses of action:
A) Subcontracting
B) Construct new facilities
C) Do nothing (no change)
The correct choice depends largely upon demand, which may be
low, medium, or high. By consensus, management estimates the
respective demand probabilities as 0.1, 0.5, and 0.4.
9
Glass Factory Example: Payoff Table
The management also estimates the profits when choosing from the
three alternatives (A, B, and C) under the differing probable levels
of demand. These profits, in thousands of dollars are presented in
the table below:
0.1 0.5 0.4
Low Medium High
EV
A 10 50 90
62
B -120 25 200 80.5
C 20 40 60 46
EVB = 0.1(-120) + 0.5(25) + 0.4(200) = 80.5 10
Step 1: 3 Decisions
A
B
0.1 0.5 0.4
Low Medium High
A 10 50 90
B -120 25 200
C 20 40 60 11
Step 2: States of Nature, Probabilities and Payoffs
Event
High demand (0.4) $90k
Medium demand (0.5) $50k
Low demand (0.1) $10k
A High demand (0.4) $200k
B Medium demand (0.5) $25k
Low demand (0.1) -$120k
C
High demand (0.4) $60k
Decision Point Medium demand (0.5) $40k
Low demand (0.1) $20k
0.1 0.5 0.4
Low Medium High
A 10 50 90
B -120 25 200
C 20 40 60 12
Step 3: Expected Value for each Decision
High demand (0.4) $90k
Medium demand (0.5) $50k
Low demand (0.1) $10k
0.1 0.5 0.4
Low Medium High
A 10 50 90
B -120 25 200
C 20 40 60 13
Step 4: Decision
High demand (0.4) $90k
$62k Medium demand (0.5) $50k
Low demand (0.1) $10k
A High demand (0.4) $200k
$80.5k $25k
B Medium demand (0.5)
Low demand (0.1) -$120k
C
High demand (0.4) $60k
$46k
Medium demand (0.5) $40k
Low demand (0.1) $20k
14
Decision Tree Method Summary
• Possible future conditions or
states of nature
• Probability of each future
condition
• Decision alternatives
• Outcome or payoff for each
alternative under every future
condition
• Decision criterion
15
Example: Hackers’ Computer Store
Due to increasing sales over the past
years, the management is considering
several options:
A. Move to new location
B. Expand store
C. Do nothing now, consider
expansion one year later
Besides your decision, the cost and revenue also depends upon
the future demand growth, which may be strong or weak. You
estimate that
n Probability of strong demand = 0.55
n Probability of weak demand = 0.45
16
Payoff Table
Alternatives Revenue Cost Profit
Move to new location, strong 195,000 x 5 yrs 210,000 765,000
Move to new location, weak 115,000 x 5 210,000 365,000
Expand store, strong 190,000 x 5 87,000 863,000
Expand store, weak 100,000 x 5 87,000 413,000
Do nothing now, expand next 170,000 x 1+ 87,000 843,000
year if strong 190,000 x 4
Do nothing now, do not 170,000 x 5 0 850,000
expand next year if strong
Do nothing now, weak 105,000 x 5 0 525,000
17
18
Decision Tree
Strong demand (0.55) $765k
$585k
Weak demand (0.45) $365k
Move to new location
Strong demand (0.55) $863k
$660.5k
Expand store
Weak demand (0.45) $413k $843k
Do nothing now Expand store
Strong demand (0.55)$850k
$703.75k
Not Expand
$850k
Weak demand (0.45) $525k
19
Value of Perfect Information:
Glass Factory Example
• Suppose we know the future demand before
decision making
– If demand=low, choose C
– If demand=medium, choose A
– If demand=high, choose B
0.1 0.5 0.4 Expected
Low Medium High Value
A 10 50 90 62
B -120 25 200 80.5
C 20 40 60 46
Best 20 50 200 107
20
Hackers’ Computer Store Example
• If you know that the demand will definitely be strong,
– The best decision is:
– Your profit is:
• If you know that the demand will definitely be weak
– Your best decision is:
– Your profit is:
21
Hackers’ Computer Store Example
• If you know that the demand will definitely be strong,
– The best decision is: Expand
– Your profit is: 863000
• If you know that the demand will definitely be weak
– Your best decision is: Do nothing
– Your profit is: 525000
22
Value of Perfect Information
• Hackers’ Computer Store Example:
VPI = 863,000*0.55 + 525,000*0.45 – 703,750 = 7150
• Glass Factory Example:
VPI = 107– 80.5 = 26.5
23
Takeaways
• Capacity Planning via Decision Tree Method
• Value of Perfect Information
• Next Class: Resource Management via Linear
Programming
24
Appendix(Optional): Math of Value of
Perfect Information
26
