Linear Programming Problem Formulation

Sasadhar Bera, IIM Ranchi

Product Mix Problem

A company wishes to schedule the production of a kitchen
appliance that requires two resources, labor and material.
The company is considering 3 models (A, B, and C) and its
production engineering department has furnished the
data given below. Formulate the following problem.

Resource
Labour (Hrs/Unit)
Material (Kg/Unit)
Profit (Rs. /Unit)

Resource
requirement
A
B
C
7
3
6
4
4
5
4
2
3

Availability
150 Hrs
200 Kg

2
Sasadhar Bera, IIM Ranchi

Product Mix Problem (contd.)

Decision Variables
XA = Number of A type model produced
XB = Number of B type model produced
XC = Number of C type model produced
Objective function: Total profit maximization (ZMAX.)
ZMax. = 4XA +2XB +3XC
Objective function
subject to
7XA +3XB +6XC 150
Labour constraint
4XA +4XB +5XC 200
Material constraint
XA , XB, XC 0
Boundary Constraint

Sasadhar Bera, IIM Ranchi

Linear Programming (LP) Formulation

Example:
Primal Problem:
Cost or profit coefficients
Technological coefficients

ZMax. = 4 X1 +2X2 +3X3

subject to

Resources

7 X1 +3 X2 +6 X3
4 X1 +4 X2 +5X3

150
200

X1 , X2, X3 0
Decision variables: X1, X2, X3
Sasadhar Bera, IIM Ranchi

How to Formulate a Linear Programming

Problem?
The representation of a problem in mathematical form
is called problem formulation. A linear Programming
(LP) must have both linear objective function and
linear constraints.

The variables which are used to formulate a linear

programming (LP) are called decision variables. They
represent the solution of the problem.
The maximization or minimization function is called the
objective function.
5
Sasadhar Bera, IIM Ranchi

How to Formulate a Linear Programming

Problem? (Contd.)
The constraints written after objective function are
called resource constraints or technological
constraints.
A sign restriction (either non-negativity or unrestricted)
is also.

6
Sasadhar Bera, IIM Ranchi

How to Formulate a Linear Programming

Problem? (Contd.)
The problem formulation has the following steps:
i.

Identifying the decision variables

ii. Writing the objective function

iii. Writing the constraints
iv. Writing the each decision variable restriction

7
Sasadhar Bera, IIM Ranchi

What is Linear Programming Model?

The word Programming in Linear programming has nothing
to do with computer programming. It originated with the
British term Programme, which is essentially a plan or a
schedule of operations.

Linear programming model is an subset of a larger class of

mathematical programming models.
Any mathematical programming model either minimize or
maximize the objective function subject to a set of
constraints.

8
Sasadhar Bera, IIM Ranchi

Inspection Problem
A company has two grades of inspector A and B that are to
be assigning to a QC inspection. It is required that at least 800
pieces be inspected per 8 hrs day.
Grade A inspector can check pieces at the rate of 25/hr with
an accuracy of 98%.
Grade B inspector can check pieces at the rate of 15/hr with
an accuracy of 95%.
Wage rate of grade A inspector Rs. 4/hr and that of grade B
inspector Rs. 3/hr.
Each time an error is made by the cost to the company Rs. 2.
The company has available for inspector 8 numbers of grade A
and 10 numbers of Grade B.
The company wants to determine the optimal assign of
inspector that minimizes total cost of inspection.
Sasadhar Bera, IIM Ranchi

Inspection Problem (contd.)

Objective function: Total cost minimization (ZMin.)
XA = Number of grade A inspectors.
XB = Number of grade B inspectors.
Company incurred two type of cost for an inspection
i) Wages cost
ii) Cost of inspection error
Hourly cost of grade A inspector = (4+25*0.02*2) = 5
Hourly cost of grade B inspector = (3+15*0.05*2) = 4.5
Zmin. = 8(5XA +4.5XB)

Objective function

Subjected to the following constraints

XA 8
Labour or resource constraint
XB 10
(25XA + 15XB)*8 800

Production requirement

XA , XB 0 and integer
Sasadhar Bera, IIM Ranchi

10

Call Center Scheduling Problem

A telephone service company maintains a 24-hour telephone answering
service. Personnel reports at 4 hours intervals and works 6 hour shifts. An
individuals wage per shift depends on the shifts starting time and given in
the table below with the minimum numbers of personnel that the
company feels must be on duty to answer calls during the six daily 4 hours
periods.
Working Shift
8 AM 2 PM

Wage
(Rs/Hr)

1900

1
Hours

Persons
Needed

2 PM 8 PM

1700

8 PM 2 AM

2 AM 8 AM

1800

2100

Period Number (Answering Service)

2
3
4
5

8 AM
12Noon

12 Noon
4 PM

4 PM
8 PM

8 PM
Mid Night

Mid Night
4 AM

4 AM
8 AM

20

12

30

15

Company is interested to run the answering service with minimum labour

cost.
11
Sasadhar Bera, IIM Ranchi

Call Center Scheduling Problem (contd.)

Xi = Number of person present at shift i .
i =1, 2, 3, 4
There are total 4 shifts. Hence total number of variables are 4. Note that a shift may cover
two periods also.

Wage
(Rs./Hr)

1900

1700

1800

2100

Working Shift

Period Number

8 AM 2 PM 2 PM 8 PM 8 PM 2 AM 2 AM 8 AM
1

8 AM 12 Noon

12 Noon 4 PM

4 PM 8 PM

8 PM Mid Night

Mid Night 4 AM

4 AM 8 AM

1
1
0
0
0
0

0
1
1
0
0
0
Sasadhar Bera, IIM Ranchi

0
0
0
1
1
0

0
0
0
0
1
1

Persons
Needed

20
12
30
15
4
7
12

Call Center Scheduling Problem (contd.)

Xi = Number of personnel presents at shift i. i =1, 2, 3, 4 .
Hence total variables are 4. Note that there are total 6 periods of answering service. We
have to find out how many people should report at the beginning of each period.

Objective function: Total cost minimization (Zmin.)

Zmin. = 6(1900 X1 + 1700X2 + 1800X3 + 2100X4)
Here the constraint is of the form: Number at work Number needed. To express these in
terms of the six constraints, we need only that an individual who starts working at the
beginning of any period works during that period and the next period.
X1
X1 + X2
X2
X3
X3 + X4
X4

20 (For Period 1)
12 (For Period 2)
30 (For Period 3)
15 (For Period 4)
4
(For Period 5)
7
(For Period 6)

X1, X2, X3, X4 0, integer

Sasadhar Bera, IIM Ranchi

13

Bank Loan Problem

A bank with full service facility formulating a loan policy involving a
maximum of 12 crores. The following table provides the pertinent data
about the different types of loans the bank deals with:
Type of loan
Personal
Car
Home
Farm
Commercial

Interest rate
0.140
0.130
0.120
0.125
0.100

Probability of bad debt

0.10
0.07
0.03
0.05
0.02

Bad debts are uncoverable and hence produce no interest revenue.

Competition with other financial institutions in the area requires the bank
to allocate at least 40% of the funds to farm and commercial loans. To assist
the housing industry in the region, home loans must equal at least 50% of
the personal, car and home loans. The bank also has a stated policy
specifying that the overall ratio for bad debts on all loans may not exceed
14
Sasadhar
Bera,
IIM
Ranchi
0.04.

Bank Loan Problem (contd.)

Objective function: Net return maximization (ZMax.)
X1 = Personal Loan in Crore
X2 = Car loan in Crore
X3 = Home loan in Crore
X4 = Farm Loan in Crore
X5 = Commercial loan in Crore
Net Return = Revenue from interest Lost fund from bad debt
ZMax.=[0.14 (0.9 X1) + 0.13 (0.93 X2) + 0.12(0.97 X3) + 0.125(0.95X4) +
0.10(0.98X5)]
Subject to
X1 + X2 + X3 + X4 + X5 12
X4 + X5 0.40*(X1 + X2 + X3 + X4 + X5)
X3 0.50 (X1 +X2 +X3)
[(0.10X1 + 0.07X2 + 0.03X3 + 0.05X4 + 0.02X5)/(X1 + X2 + X3 + X4 + X5)] 0.04
X1, X2, X3, X4, X5 0
Sasadhar Bera, IIM Ranchi

15

Measuring Restaurants Efficiency

Pradip Nair owns three fast food restaurants located in three different
places in Ranchi city. Nair is interested to identify most efficient restaurant
and wants to improve the efficiency of other relatively less efficient
restaurants. Nair has set three inputs and three outputs factor to measure
the efficiency of a restaurant. The relevant information is given below.
A
Input Measures
Weekly hours of operation
(x1)
Daily number of employees
(x2)
Daily cost of ingredients
(x3)
Output Measures
Weekly profit (y1)
Number of customer
served/week (y2)
City fast food market share
(y3)

Restaurants
B

54

40

58

2000

4000

3000

5000
1500

4000
1000

3500
1200

30%

20%

15%
16

Sasadhar Bera, IIM Ranchi

Measuring Restaurants Efficiency (contd.)

We have to compute the efficiency for three different restaurants. Hence we
have to formulate the three separate linear programming (LP)

Objective function: Maximize weighted sum of output measures (ZMax.)

Formulation for Restaurant A:
ZMax.= v1*y1+ v2*y2+v3*y3 = v1*5000 + v2*1500 + v3*0.30
v1, v2, v3 weight associated with output measures y1, y2, y3 respectively.

Subject to

(weighted sum of output)/(weighted sum of input) 1

weighted sum of output - weighted sum of input 0

v1*5000 + v2*1500 + v3*0.30 - u1*54 + u2*6 + u3*2000 0

v1*4000 + v2*1000 + v3*0.20 - u1*40 + u2*8 + u3*4000 0
v1*3500 + v2*1200 + v3*0.15 - u1*58 + u2*7 + u3*3000 0
u1, u2, u3 weight associated with input measures x1, x2, x3 respectively.
To ensure a bounded solution the sum of weighted inputs should be 1.
u1*54 + u2*6 + u3*2000 = 1
u1, u2, u3, v1, v2, v3 0
Sasadhar Bera, IIM Ranchi

17

Measuring Restaurants Efficiency (contd.)

Formulation for Restaurant B:
ZMax = v1*4000 + v2*1000 + v3*0.20
Subject to

v1*5000 + v2*1500 + v3*0.30 - u1*54 + u2*6 + u3*2000 0

v1*4000 + v2*1000 + v3*0.20 - u1*40 + u2*8 + u3*4000 0
v1*3500 + v2*1200 + v3*0.15 - u1*58 + u2*7 + u3*3000 0
u1*40 + u2*8 + u3*4000 = 1
u1, u2, u3, v1, v2, v3 0

Sasadhar Bera, IIM Ranchi

18

Measuring Restaurants Efficiency (contd.)

Formulation for Restaurant C:

ZMax = v1*3500 + v2*1200 + v3*0.15

Subject to
v1*5000 + v2*1500 + v3*0.30 - u1*54 + u2*6 + u3*2000 0
v1*4000 + v2*1000 + v3*0.20 - u1*40 + u2*8 + u3*4000 0
v1*3500 + v2*1200 + v3*0.15 - u1*58 + u2*7 + u3*3000 0
u1*58 + u2*7 + u3*3000 = 1
u1, u2, u3, v1, v2, v3 0

Sasadhar Bera, IIM Ranchi

19

Multi-period Scheduling Problem

A plastic molding manufacturer is planning the production, storage and marketing
of a product whose demand and selling price vary seasonally. The company is in
the process of developing a manufacturing strategy. According to the formulated
strategy, the company wants to manufacture the product in the season during
which production cost are low and store and sell the product when the selling price
is high. The manufacturer has obtained the following estimates of the demand,
cost, capacity, and expected selling price of the product for each period. Consider
beginning inventory 500 tons and desired ending inventory 300 tons, storage cost
(i. e. inventory holding cost) is Rs. 2 per ton from one period to the next.
Period
1
2
3
4
5
6

Production
Costs (Rs.)
20
25
30
40
50
60

Production
Capacity (ton)
1500
2000
2200
3000
2700
2500

Demand (tons)
(D)
1100
1500
1800
1600
2300
2500

Selling Price
(Rs.)
180
180
250
270
300
320

Formulate the linear programming problem that provides the optimal production
Sasadhar Bera, IIM Ranchi
20
schedule and maximize profit.

Multi-period Scheduling Problem (contd.)

In multi-period production scheduling total number of period 6. We have
denoted it by j where j = 1, 2, 3 . . .,6
Xj = Product made in tons during period j
Dj = Product demanded during period j
Ij = Product left in storage (called inventory) at the end of period j
In this formulation problem, first we write the constraints and next objective
function.
Beginning inventory 500 tons. Hence first constraint
i) I0 = 500
Production at period 1 + beginning inventory demand at period 1
ii) X1 + I0 D1
Production at period 2 + inventory at period 1 demand at period 2
iii) X2 + I1 D2
Similarly for other periods
iv) X3 + I2 D3 v) X4 + I3 D4 vi) X5 + I4 D5 vii) X6 + I5 D6

21
Sasadhar Bera, IIM Ranchi

Multi-period Scheduling Problem (contd.)

Inventory (which is the extra amount that is produced) at period 1
viii) I1 = X1 + I0 D1
Inventory at period 2:
ix) I2 = X2 + I1 D2

Similarly for other periods

x) I3 = X3 + I2 D3
xi) I4 = X4 + I3 D4
xii) I5 = X5 + I4 D5
xiii) I6 = X6 + I5 D6
desired ending inventory 300 tons. Hence inventory at period 6
xiv) I6 = 300

Sasadhar Bera, IIM Ranchi

22

Multi-period Scheduling Problem (contd.)

Others constraints related to capacity and demand:
X1 1500, X2 2000, X3 2200, X4 3000, X5 2700, X6 2500
Variable boundary constraints
X1, . . .,X6 , I1, . . ., I6 0

Objective function: Maximize profit (ZMax.)

Profit = (selling price cost of production) inventory cost

ZMax = (180-20)*X1 + (180-25)*X2 + (250-30)*X3 + (270-40)*X4
+ (300-50)*X5 + (320-60)*X6 2*(I1+I2+I3+I4+I5+I6)

Sasadhar Bera, IIM Ranchi

23