Michael K. answered 04/26/19
PhD professional for Math, Physics, and CS Tutoring and Martial Arts
Given this matrix (call it A) we are looking to solve for the null space of A...
A = | 2 2 0 |
| P 2 0 |
| Q -1 -1 |
We can use row-echelon formalism to reduce the matrix A into a upper triangular format for easy solution gathering...
1) P*R1 - 2*R2 --> R2
A = | 2 2 0 |
| 0 (2P-4) 0 |
| Q -1 -1 |
2) Q*R1 - 2*R3 --> R3
A = | 2 2 0 |
| 0 (2P - 4) 0 |
| 0 (2Q - 2) -1 |
3a) The final representation is to get 2Q - 2 = -(2P - 4). We can write one of the unknowns (at this time) in terms of the other.
2Q - 2 = 4 - 2P
2Q = 6 - 2P
Q = 3 - P
3b) R2 + R3 --> R3
A = | 2 2 0 |
| 0 (2P - 4) 0 |
| 0 (2P+2Q - 6) -1 |
A = | 2 2 0 |
| 0 (2P - 4) 0 |
| 0 2(P + Q - 3) -1 |
With the constraint above (in bold) we finally get...
A = | 2 2 0 |
| 0 (2P - 4) 0 |
| 0 0 -1 |
Therefore from R3 we see that
-1 * Z = 0 --> Z = 0
From R2 we see that
(2P - 4)Y = 0 --> Y = 0 or (2P - 4) = 0, Since we need to solve for P, we requested that Y ≠ 0 which leads to P = 2 and therefore Q = 1
From R1 we see that
2X + 2Y = 0 --> since Y is arbitrary. we can define X in terms of Y --> X = -Y
Finally our matrix and solution vector can be seen as
A = | 2 2 0 |
| 2 2 0 |
| 1 -1 -1 |
With solution vector (parametrized) of s * (-1, 1, 0) where s can be any value.