Linear Algebra - Complete solution for Ax = b

2.7k Views Asked by At

Alright, I'm having some trouble understanding the "complete" solution for Ax = b.

For instance, suppose

$$A = \pmatrix{ 1 & 2 & 2 & 2 \\ 2 & 4 & 6 & 8 \\ 3 & 6 & 7 & 10}$$

I can already see that $Ax = \pmatrix{1 \\ 5 \\ 6}$ is a solution for this system but after elimination we get

$$A = \pmatrix{ 1 & 2 & 2 & 2 &b_1\\ 0 & 0 & 2 & 4 & b_2 - 2b_1 \\ 0 & 0 & 0 & 0 & b_3 - b_2 - b_1 0}$$

From there my textbook shows how to find a particular solution by setting all free variables to 0, yielding:

$$ x_1+ 2x_3 = 1$$ $$2x_3 = 3 $$

So that

$$x-particular = \pmatrix{-2 \\ 0 \\ \frac{3}{2} \\ 0}$$

It then claims that the complete solution to Ax = b is given by $$x-complete + x_n$$ where $x_n$ is a "generic vector in the nullspace",and since $$x_n = c_1 * \pmatrix{-2 \\ 1 \\ 0 \\ 0} + c_2 * \pmatrix{2 \\ 0 \\ -2 \\ 1}$$

$$x-complete = \pmatrix{-2 \\ 0 \\ \frac{3}{2} \\ 0} + c_1 * \pmatrix{-2 \\ 1 \\ 0 \\ 0} + c_2 * \pmatrix{2 \\ 0 \\ -2 \\ 1}$$

What I don't understand is:

  • what is this "generic vector in the nullspace?
  • Do we have to set the free variables to 0 and 1 respectively in order to find the nullspace? If not, how can this solution be considered complete if we're not describing all the possibles values the free variables could've taken?
1

There are 1 best solutions below

0
On BEST ANSWER

Since $m>n$ the system $Ax=b$ has infinitely many or zero solutions depending upon the augmented RREF

$$A = \pmatrix{ 1 & 2 & 2 & 2 &b_1\\ 0 & 0 & 2 & 4 & b_2 - 2b_1 \\ 0 & 0 & 0 & 0 & b_3 - b_2 - b_1 }$$

Notably if $b_3-b_2-b_1\neq 0$ we have no solution otherwise the general solution is given by $x_P+x_H$ that is the sum of

  • one particular soution to $Ax_P=b$
  • the homogeneous solution to $Ax_H=0$