Linear programming: Examples

June 20, 2021

1/8
In linear programming problem, we try to find the best outcome of a function
that is a linear combination of the unknown variables with constraints that are
also linear in the unknown variables.

This is a linear function in x1 , x2 , ..., xN :

a1 · x1 + a2 · x2 + . . . + aN · xN (1)

The solution of the unknown variables should satisfy the constraints. An

equality constraint
x1 + x2 = 0 (2)
means that the unknown variables x1 + x2 must be equal to 0. An inequality
constraint
x1 ≥ 0 (3)
means that x1 should be larger than or equal to 0.

2/8
For a two variable problem, the feasible region has the following form for the
considered constraints:

2
4
x2 3
Feasible
region

x1
3/8
Find the maximum of the following function:

z = 3 · x1 + 2 · x2 (4)

subjected to

1 · x1 + 1 · x2 ≤ 15 (5)
4
· x1 + 1 · x2 ≤ 16 (6)
3
2 · x1 + 1 · x2 ≤ 20 (7)

(This is a high school maths problem that you can already solve).

4/8
The objective function we want to maximize:

z = 3 · x1 + 2 · x2 (8)

20
75
10 60
45
30
x2

0
15
0
−10 −15
−30
−20
0 5 10 15
x1

5/8
The different constraints are presented on separate graphs (a) - (c):

(a) (b)
20 20
x2 = 15 − x1 75 x2 = 16 − 4/3x1 75
10 60 10 60
45 45
30 30
x2

x2
0 0
15 15
0 0
−10 −15 −10 −15
−30 −30
−20 −20
0 5 10 15 0 5 10 15
x1 x1

(c) (d)
20 20
x2 = 20 − 2x1 75 x2 = (10 − 3 · x1 )/2 75
10 60 10
x2 = (20 − 3 · x1 )/2 60
x2 = (34 − 3 · x1 )/2
45 45
30 30
x2

x2
0 0
15 15
0 0
−10 −15 −10 −15
−30 −30
−20 −20
0 5 10 15 0 5 10 15
x1 x1

Three lines corresponding to distinct level sets of the objective function are
presented in (d).

6/8
The graphical solution of the linear programming problem:

10
x2 = 15 − x1 47.85
x2 = 16 − 4/3x1 45.30
x2 = 20 − 2x1
42.75
x2 = (34 − 3 · x1 )/2
8 40.20
37.65
x2

35.10
32.55
6 30.00
27.45
24.90
5 6 7 8 9 10
x1

Optimal variables:
x1∗ = 6 and x2∗ = 8 (9)
The function value corresponding to the optimal variables:

z ∗ = 34 (10)

7/8
We can easily do the same in Excel:

SUMPRODUCT is very useful.

8/8