Session # 6

Integer programming

l
ia
er
at
20 M
9- rse
P1 u
M I Co
,G R
M L
M X
Course: Operations/ Optimization Models for

O of
Shounak Basak
O rt
Managerial Decisions Pa
Room no. 6, Library Building,
Programme: GMP19 XLRI – Xavier School of Management,
Term II, 2019 Jamshedpur – 831001
Jamshed1
Phone: 0657 – 665 – 3424
Email: [email protected]

Shounak Basak
PODS Area, XLRI – Xavier School of Management
 l
ia
er
at
20 M
9- rse
Integer programming

P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
2
Types of integer programming

• Pure Integer programming problems

l
ia
er
at
• Mixed Integer programming problems

20 M
9- rse
P1 u
• Binary Integer programming problems

M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
3
LP sensitivity problem
Problem # 6.1

l
ia
er
Solve the following problem graphically:

at
max 𝑧 = 5𝑥1 + 10𝑥2

20 M
Subject to −2𝑥1 + 4𝑥2 ≤ 6

9- rse
6𝑥1 + 3𝑥2 ≤ 30

P1 u
M I Co
𝑥1 , 𝑥2 ≥ 0; 𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
4
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
5
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
6
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
7
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
8
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
9
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
10
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
11
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
12
Solution
Problem # 6.1

max 𝑧 = 5𝑥1 + 10𝑥2

Subject to −2𝑥1 + 4𝑥2 ≤ 6

l
ia
6𝑥1 + 3𝑥2 ≤ 30

er
𝑥1 , 𝑥2 ≥ 0;

at
20 M
𝑥1 , 𝑥2 𝑎𝑟𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟𝑠

9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
13
Binary integer programming

• Either – or alternatives:

l
ia
er
at
20 M
If either 𝑥1 or 𝑥2 are being considered, then 𝑥1 + 𝑥2 = 1, with 𝑥1 , 𝑥2 𝑏𝑒𝑖𝑛𝑔 0,1

9- rse
P1 u
If neither of them is also a possibility, then 𝑥1 + 𝑥2 ≤ 1

M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
14
Binary integer programming

• 𝑘 – out of – 𝑛 alternatives:

l
ia
er
at
If 2 machines out of 5 machines are to be active: 𝑥1 + 𝑥2 + 𝑥3 + 𝑥4 + 𝑥5 = 2

20 M
9- rse
P1 u
If at least 2 machines out of 5 machines are to be active: 𝑥1 + 𝑥2 + 𝑥3 + 𝑥4 + 𝑥5 ≥ 2

M I Co
,G R
If at most 2 machines out of 5 machines are to be active: 𝑥1 + 𝑥2 + 𝑥3 + 𝑥4 + 𝑥5 ≤ 2

M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
15
Binary integer programming

l
ia
er
• 𝑖𝑓 − 𝑡ℎ𝑒𝑛 alternatives:

at
20 M
9- rse
If 𝑥2 requires 𝑥1 to occur but not vice versa: 𝑥1 − 𝑥2 ≥ 0

P1 u
M I Co
If both 𝑥1 and 𝑥2 require the other to occur: 𝑥1 − 𝑥2 = 0

,G R
M L
M X
O of
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
16
Binary integer programming

• Either – or constraints:

• For lower limit constraints:

l
ia
er
at
2𝑥1 + 3𝑥2 ≥ 20 , 4𝑥1 + 5𝑥2 ≥ 30

20 M
The either – or constraint is:

9- rse
P1 u
M I Co
2𝑥1 + 3𝑥2 ≥ 20𝑦1 , 4𝑥1 + 5𝑥2 ≥ 30(1 − 𝑦1 )

,G R
Where, 𝑦1 is binary variable

M L
M X
O of
• For upper limit constraints:
O rt
Pa
2𝑥1 + 3𝑥2 ≤ 20 , 4𝑥1 + 5𝑥2 ≤ 30
The either – or constraint is:

2𝑥1 + 3𝑥2 ≤ 20 +𝑀𝑦1 , 4𝑥1 + 5𝑥2 ≤ 30 + 𝑀(1 − 𝑦1 )

Where, 𝑦1 is binary variable and 𝑀 is a very large number
Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
17
Binary integer programming

• Minimum level requirements:

l
ia
er
at
If 𝑥1 is greater than 200 when selected:

20 M
9- rse
𝑦1 is 1 when selected,

P1 u
M I Co
Then, 𝑥1 − 200𝑦1 ≥ 0, where, 𝑦1 is binary

,G R
M L
M X
O of
𝑥1 ≤ 𝑀𝑦1
O rt
Pa

Shounak Basak
9/13/2019
PODS Area, XLRI – Xavier School of Management
18
References
Material for the Lecture has been sourced from: Selected References:

Stevenson, W. J., & Ozgur, C. (2006). Introduction to management science with

spreadsheets and student CD. McGraw-Hill, Inc...
Hillier, F. S., & Lieberman, G. J. (2012). Bodhibrata Nag and Preetam

l
ia
Basu,“Introduction to Operations Research”.

er
Vanderbei, R. J. (2015). Linear programming. Heidelberg: Springer.

at
G. Hadley, Linear Programming, Addison-Wesley, Reading, Massachusetts, 1962

20 M
9- rse
P1 u
M I Co
,G R
M L
M X
O of
O rt
Pa

Shounak Basak
13-09-2019
PODS Area, XLRI – Xavier School of Management
19