In this section, we provide an example of Quadratic Programming. We will use the simplex method to solve this problem.
Maximize f(x) = 2x1 + 3x2 x12 x22
subject to
x1 + x2 ≤ 2
2x1 + x2 ≤ 3
x1, x2 ≥ 0.
Solution.
First, find that a given function is concave or convex.
df(x)/dx1 = 2 2x1
df(x)/dx2 = 3 2x2
d2f(x)/dx12 = 2
d2f(x)/dx22 = 2
d2f(x)/dx1.dx2 = 0
H(x) = | d2f(x)/dx12 | d2f(x)/dx1.dx2 |
d2f(x)/dx1.dx2 | d2f(x)/dx22 |
H(x) = | 2 | 0 |
0 | 2 |
I = | 1 | 0 |
0 | 1 |
λ(I) | λ | 0 |
0 | λ |
H(x) λ(I) | 2 | 0 | | λ | 0 |
0 | 2 | 0 | λ |
H(x) λ(I) | 2 λ | 0 |
0 | 2 λ |
Determinant of H(x) λ(I) = 0
( 2 λ ) X ( 2 λ ) 0 X 0 = 0
λ2 + 4λ + 4 = 0
λ2 + 2λ + 2λ + 4 = 0
λ(λ + 2) + 2(λ + 2) = 0
λ = 2
Since λ is negative, the given problem has a concave objective function.
φ (x, λ ) = 2x1 + 3x2 x12 x22 + λ1(2 x1 x2) + λ2(3 2x1 x2)
Differentiate w.r.t. x1
φ x1 = 2 2x1 λ1 2λ2 = μ1.....(i)
Differentiate w.r.t. x2
φ x2 = 3 2x2 λ1 λ2
= μ2.....(ii)
Where μ1 and μ2 are surplus variables.
x1, x2, λ1, λ2, μ1, μ2 ≥ 0
μ1x1 = 0, μ2x2 = 0
Differentiate w.r.t. λ 1
φ λ1 = 2 x1 x2 = S1.....(iii)
Differentiate w.r.t. λ 2
φ λ2 = 3 2x1 x2 = S2.....(iv)
Where S1 and S2 are slack variables.
λ1S1 = 0, λ2S2 = 0
From equations (i), (ii), (iii) & (iv), we get
2x1 + λ1 + 2λ2 μ1 = 2
2x2 + λ1 + λ2 μ2 = 3
x1 + x2 + S1 = 2
2x1 + x2 + S2 = 3
Introducing artificial variables A1and A2 to the first two equations to obtain a feasible solution of the problem.
Maximize A1 A2
subject to
2x1 + λ1 + 2λ2 μ1 + A1 = 2
2x2 + λ1 + λ2 μ2 + A2 = 3
x1 + x2 + S1 = 2
2x1 + x2 + S2 = 3
Now, we solve the above problem by Simplex method.
Table 1
cj | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 1 | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|
cB | Basic variables B |
x1 | x2 | μ1 | μ2 | S1 | S2 | λ1 | λ2 | A1 | A2 | Solution values b (=XB) |
1 | A1 | 2 | 0 | 1 | 0 | 0 | 0 | 1 | 2 | 1 | 0 | 2 |
1 | A2 | 0 | 2 | 0 | 1 | 0 | 0 | 1 | 1 | 0 | 1 | 3 |
0 | S1 | 1 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 2 |
0 | S2 | 2 | 1 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 3 |
zjcj | 2 | 2 | 1 | 1 | 0 | 0 | 2 | 3 | 0 | 0 |
If we ensure that the pairs ( λi,
Si) and (μj,
xj) are not basic variables simultaneously then the conditions
Siλi = 0 and xjμj = 0 will be automatically satisfied.
Although zj cj is lowest corresponding
to λ2 column, it cant be made a basic variable as S2 is already
a basic variable. Hence, A1 departs and x1 enters.
Table 2
cj | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | ||
---|---|---|---|---|---|---|---|---|---|---|---|
cB | Basic variables B |
x1 | x2 | μ1 | μ2 | S1 | S2 | λ1 | λ2 | A2 | Solution values b (=XB) |
0 | x1 | 1 | 0 | 1/2 | 0 | 0 | 0 | 1/2 | 1 | 0 | 1 |
1 | A2 | 0 | 2 | 0 | 1 | 0 | 0 | 1 | 1 | 1 | 3 |
0 | S1 | 0 | 1 | 1/2 | 0 | 1 | 0 | 1/2 | 1 | 0 | 1 |
0 | S2 | 0 | 1 | 1 | 0 | 0 | 1 | 1 | 2 | 0 | 1 |
zjcj | 0 | 2 | 0 | 1 | 0 | 0 | 1 | 1 | 0 |
Table 3
cj | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | ||
---|---|---|---|---|---|---|---|---|---|---|---|
cB | Basic Variables B |
x1 | x2 | μ1 | μ2 | S1 | S2 | λ1 | λ2 | A2 | Solution values b (=XB) |
0 | x1 | 1 | 0 | 1/2 | 0 | 0 | 0 | 1/2 | 1 | 0 | 1 |
1 | A2 | 0 | 0 | 1 | 1 | 2 | 0 | 2 | 3 | 1 | 1 |
0 | x2 | 0 | 1 | 1/2 | 0 | 1 | 0 | 1/2 | 1 | 0 | 1 |
0 | S2 | 0 | 0 | 1/2 | 0 | 1 | 1 | 1/2 | 1 | 0 | 0 |
zjcj | 0 | 0 | 1 | 1 | 2 | 0 | 2 | 3 | 0 |
Since S2 is a basic variable, λ2 cant be made a basic variable.
Therefore, the variable A2 departs and λ1 enters.
cj | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
---|---|---|---|---|---|---|---|---|---|---|
cB | Basic variables B |
x1 | x2 | μ1 | μ2 | S1 | S2 | λ1 | λ2 | Solution values b (=XB) |
0 | x1 | 1 | 0 | 1/4 | 1/4 | 1/2 | 0 | 0 | 1/4 | 3/4 |
0 | λ1 | 0 | 0 | 1/2 | 1/2 | 1 | 0 | 1 | 3/2 | 1/2 |
0 | x2 | 0 | 1 | 1/4 | 1/4 | 1/2 | 0 | 0 | 1/4 | 5/4 |
0 | S2 | 0 | 0 | 1/4 | 1/4 | 3/2 | 1 | 0 | 1/4 | 1/4 |
zjcj | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
The values for x1 and x2 are 3/4
and 5/4 respectively.
The associated optimal value of the objective function is f(x) = 2 X
3/4 + 3 X 5/4 (3/4)2 (5/4)2 = 25/8