LINEAR PROGRAMMING
CASE STUDIES
Vassilis Kostoglou
E-mail: [email protected]
URL: www.it.teithe.gr/~vkostogl
1
LINEAR PROGRAMMING – CASE STUDIES
Case study 1
Portfolio selection
Select a portfolio package from a set of alternative investments
Maximization of the expected return or minimization of the risk
Available capital
Company’s policy
Duration of investments’ economic life, potential growth rate, danger, liquidity
2
LINEAR PROGRAMMING – CASE STUDIES
Return data
Expected annual return of investments
Investment Expected annual return rate (%)
Share A – manufacturing sector 15.4
Share B - manufacturing sector 19.2
Share C - food and beverage sector 18.7
Share D – food and beverage sector 13.5
Mutual fund E 17.8
Mutual fund Z 16.3
3
LINEAR PROGRAMMING – CASE STUDIES
Requirements
Total amount = € 90000
Amount in shares of a sector no larger than 50% of total available
Amount in shares with the larger return of a sector less or equal to 80% of sector’s
total amount
Amount in manufacturing company Β less or equal to 10% of the whole share
amount
Amount in mutual funds less or equal to 25% of the amount in manufacturing
shares
4
LINEAR PROGRAMMING – CASE STUDIES
Solution
Decision variables
x1 = invested amount in share A of the manufacturing sector
x2 = invested amount in share B of the manufacturing sector
x3 = invested amount in share C of the food and beverage sector
x4 = invested amount in share D of the food and beverage sector
x5 = invested amount in mutual fund E
x6 = invested amount in mutual fund Z
5
LINEAR PROGRAMMING – CASE STUDIES
Case Study 2
Financial programming problem
Initial amount: € 80000
Timeframe of investments’ decisions: 4 months
Two-month government bonds: return 3%
Three-month government bills: return 6.5%
Bank deposits: interest 1%
At the beginning of the 5th month at least € 40000 are needed
Maximum amount in two-month or three-month bonds: € 32000
6
LINEAR PROGRAMMING – CASE STUDIES
Decision variables
Βj = amount to be invested in government bonds at the beginning of the month j
Cj = amount to be invested in government bills at the beginning of the month j
Dj= amount to be invested in bank deposits at the beginning of the month j
7
LINEAR PROGRAMMING – CASE STUDIES
Case Study 3
Investment choice problem in a limited capitals status
There are five independent investments
Maximization of the total present net value
Satisfaction of budget limitations
Cash inflows of the investments
Cash outflows of the investments
Each investment is divisible (investment rate)
8
LINEAR PROGRAMMING – CASE STUDIES
Inflow data
Cash inflows
Year Investment
1 2 3 4 5
1 38 11 17 8 25
2 41 16 24 11 28
3 54 15 29 13 35
4 - 20 - 19 46
9
LINEAR PROGRAMMING – CASE STUDIES
Outflow data
Cash outflows
Year Investment Amount
1 2 3 4 5 Available
0 34 10 16 9 31 55
1 13 5 8 4 10 28
2 14 6 10 6 13 30
3 17 6 11 7 12 37
4 - 8 - 5 16 30
10
LINEAR PROGRAMMING – CASE STUDIES
Case study 4
Advertising media selection
Problem data
Advertising media Cost of one view Units of expected audience
(in €) rate of one view
1. Friday – day 400 5000
2. Saturday – day 450 5500
3. Sunday – day 450 5700
4. Friday – night 500 7500
5. Saturday – night 550 8200
6. Sunday - night 550 8400
11
LINEAR PROGRAMMING – CASE STUDIES
Other relative elements
Goal: Determination of views / records in order to maximize the total audience rate
Total available amount: € 45000
Maximum amount for Friday: € 11000
Maximum amount for Saturday: € 14400
Total daily view number: at least 20
Total nightly view number: at least 50% of the total
Maximum view number: each day 12, each night 18
12
LINEAR PROGRAMMING – CASE STUDIES
Decision variables
Χ1 = number of views on Friday (day)
Χ2 = number of views on Saturday (day)
Χ3 = number of views on Sunday (day)
Χ4 = number of views on Friday (night)
Χ5 = number of views on Saturday (night)
Χ6 = number of views on Sunday (night)
13
LINEAR PROGRAMMING – CASE STUDIES
Case study 5
Marketing research
Personal interviews, Daily (D) and Nightly (N)
Households: with children, without children, of one person
Sample = 800 households
At least: 400 with children, 200 without children, 100 of one person
Respondents Night > Respondents Day
To be done during: At least 50% of interviews to households with children
At least 60% of interviews to households without children
At least 70% of interviews to households of one person
14
LINEAR PROGRAMMING – CASE STUDIES
Cost elements
Interview cost (in monetary units)
Household category Day Night
With children 1500 1800
Without children 1300 1600
One person 1000 1200
15
LINEAR PROGRAMMING – CASE STUDIES
Decision variables
x11 = Number of interviews to households with children carried out during the day
x12 = Number of interviews to households with children carried out during the night
x21= Number of interviews to households without children carried out during the day
x22 = Number of interviews to households without children carried out during the night
x31 = Number of interviews to households of one person carried out during the day
x32 = Number of interviews to households of one person carried out during the night
16
LINEAR PROGRAMMING – CASE STUDIES
Case study 6
Human resources management
Allocation of available human resources to different departments, work centers,
shifts etc.
Recruitment of seasonal staff
Allocation of staff to shifts
Minimizing the number of employees who should work in various time periods
during the day
17
LINEAR PROGRAMMING – CASE STUDIES
Problem’s data
Time period Shift Minimum number of
1 2 3 4 5 required employees
07 a.m. – 09 a.m. + 35
09 a.m. – 11 a.m. + + 68
11 a.m. – 13 p.m. + + 60
13 p.m. – 15 p.m. + + + 57
15 p.m. – 17 p.m. + + + 65
17 p.m. – 19 p.m. + + 63
19 p.m. – 21 p.m. + + 72
21 p.m. – 23 p.m. + 33
23 p.m. – 07 a.m. + 12
Gross employee cost 230 220 225 240 260
per day (in €)
18
LINEAR PROGRAMMING – CASE STUDIES
Case study 7
Production planning
Planning horizon: A number of time periods
Problem: Having a forecast for each period’s demand, determine the products’
quantities that can be produced with feasible methods in order to satisfy the total
demand with the minimum cost.
For two products Α and Β, there are demand forecasts for January, February and
March.
Initial stock: 100 units of product A and 120 units of product B.
Minimum total required stock: 130 units of product Α and 110 of product Β
Unit production cost: Α = € 20 and Β = € 25
Maintenance cost per period and per unit: 2% on the unit production cost
19
LINEAR PROGRAMMING – CASE STUDIES
Demand
Bicycle demand
Month Bicycle
A B
January 700 800
February 900 600
March 1000 900
Total 2600 2.300
20
LINEAR PROGRAMMING – CASE STUDIES
Capacity
System’s capacity
Month Machine capacity Available work
(machine hours) (man-hours)
January 3000 2500
February 2800 2300
March 3600 2400
21
LINEAR PROGRAMMING – CASE STUDIES
Use of resources
Required resources per product unit
Bicycle Machine hours Man-hours
A 1.5 1.1
B 1.6 1.2
22
LINEAR PROGRAMMING – CASE STUDIES
Decision variables
xIJ = number of units of product I produced during month J,
where I = A, B and J = 1, 2, 3
IIJ = number of units of product I maintained in stock at the end of the month J,
where I = A, B and J = 1, 2, 3
J = 1 – January
J = 2 – February
J = 3 – March
23
LINEAR PROGRAMMING – CASE STUDIES
Case study 8
Diet problem
Identification of a diet or of a prescription meeting specific dietary requirements
Criterion: minimum cost
Xj = the amount of ingredient j for the production of one unit of animal feed
24
LINEAR PROGRAMMING – CASE STUDIES
Problem’s data
Required nutritional Number of nutritional ingredients Nutritional requirement
ingredient per ingredient unit per animal feed unit
Ingredient
1 2 3 4
Vitamin A 80 115 100 90 >= 80
Vitamin C 110 90 85 100 >= 100
Vitamin E 50 70 105 80 >= 60
Proteins 250 300 210 240 >= 260
Calories 480 510 470 530 <= 2300
Unit cost 180 160 145 200
25
LINEAR PROGRAMMING – CASE STUDIES
Case study 9
Mix problem
Determination of the best mix program of raw material for the production of final
products
Three main ingredients Α, Β and C
Three products: super fuel, unleaded, super unleaded
Minimum required octane number
Maximization of the total daily profit
Available quantities of main ingredients
Minimum required product quantities
26
LINEAR PROGRAMMING – CASE STUDIES
Problem’s data
Main ingredient Octane number Cost per ton (€) Maximum daily
available quantity (tones)
A 120 38 1000
B 90 42 1200
C 130 105 700
27
LINEAR PROGRAMMING – CASE STUDIES
Demand data
Fuel Octane number Cost per ton (€) Daily demand (tones)
A 94 85 800
B 92 80 1100
C 96 88 500
28
LINEAR PROGRAMMING – CASE STUDIES
Decision variables
Xij = quantity of ingredient i mixed for the production of one tone of product j
for i = A, B, C and j = 1, 2, 3
j = 1 – fuel super
j = 2 – unleaded fuel
j = 3 – super unleaded fuel
29
LINEAR PROGRAMMING – CASE STUDIES
