6.1 Algorithm for solving $Ax=b$
The idea behind this exercise is to come up with the solution for $Ax=b$. Let us start by taking a matrix $A$ and vector $b$.
$$\begin{align} A = \begin{bmatrix} 1 & 2 & 2 & 2 \\ 2 & 4 & 6 & 8 \\ 3 & 6 & 8 & 10 \end{bmatrix}; b = \begin{bmatrix} b_1 \\ b_2 \\ b_3 \end{bmatrix} \end{align}$$
The augumented matrix $\begin{bmatrix} A & b \end{bmatrix}$ can be formed and elimination steps can be performed on it as follows.
$$\begin{align} A = \begin{bmatrix} 1 & 2 & 2 & 2 & b_1\\ 2 & 4 & 6 & 8 & b_2\\ 3 & 6 & 8 & 10 & b_3 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 2 & 2 & 2 & b_1\\ 0 & 0 & 2 & 4 & b_2-2b_1\\ 0 & 0 & 2 & 4 & b_3-3b_1 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 2 & 2 & 2 & b_1\\ 0 & 0 & 2 & 4 & b_2-2b_1\\ 0 & 0 & 0 & 0 & b_3-b_2-b_1 \end{bmatrix} = U \end{align}$$
The last row of the augumented matrix after elimination implies that $b_3-b_2-b_1 = 0$ and this is the condition for solvability. The condition of solvability can be listed as:
- $Ax=b$ is solvable when $b \in C(A)$
- If a combination of rows of $A$ gives a zero row, the same combination of entries of $b$ must be $0$.
One of the valid values of $b$ for which we will have the solution can be $\begin{bmatrix} 1 \\ 5 \\ 6 \end{bmatrix}$. Putting this value for $b$, we get the following augumented row-echelon matrix.
\begin{align} U = \begin{bmatrix} 1 & 2 & 2 & 2 & 1\\ 0 & 0 & 2 & 4 & 3\\ 0 & 0 & 0 & 0 & 0 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 2 & 0 & -2 & -2\\ 0 & 0 & 1 & 2 & 3/2\\ 0 & 0 & 0 & 0 & 0 \end{bmatrix} = R \end{align}
To find the complete solution of $Ax=b$, we can start with finding a particular solution $(x_p)$ and then proceed from there. The steps are as follows:
- Set all free varibales in $x$, variables corresponding to free columns to $0$ and solve for the pivot variables. This gives us the set of equations as: $x_1 + 2x_3 =1;2x_3=3;x_2=0;x_4=0$ (as $x_2,x_4$ are free variables). Hence, $x_p = \begin{bmatrix}-2 \\ 0 \\ 3/2 \\ 0 \end{bmatrix}$.
- The complete solution can be given as: \begin{align} x = x_p + x_n \end{align} where $x_n$ is any vector in the null space. This is valid as the particular solution gives $Ax_p = b$ and any vector in null space gives $Ax_n = 0$. Adding these two, we get $x=x_p+x_n$. For this particular $A$, the null space is $\begin{bmatrix} -2 & 0 \\ 2 & -2 \\ 1 & 0 \\ 0 & 1 \end{bmatrix}$. Hence the complete solution is: \begin{align} x = \begin{bmatrix} -2 \\ 0 \\ 3/2 \\ 0 \end{bmatrix}+ c_1\begin{bmatrix} -2 \\ 2 \\ 1 \\ 0 \end{bmatrix} + c_2\begin{bmatrix} 0 \\ -2 \\ 0 \\ 1 \end{bmatrix} \end{align}
The above solution forms a two-dimensional plain in $R^4$ which goes through $x_p$. It should be noted that this is not a subsapce as it does not goes through origin.
Let us further explore about the nature of the solutions of $Ax=b$. For a $m \times n$ matrix $A$, the rank $r$ satisfies: $r \leq m$ and $r \leq n$, as there can’t be more pivots than number of columns or number of rows in the matrix.
-
A full column rank means $r=n < m$. A full columnn rank means we have $n$ pivots, no free variables and hence the nullspace will only have the zero vector. It’s reduced row-echelon matrix will look like $\begin{bmatrix} I \\ 0 \end{bmatrix}$. This means that the solution $x$ for $Ax=b$ will be just $x_p$ i.e. a unique solution if it exists i.e. $0$ or $1$ solution.
-
A full row rank means $r=m < n$. This means we will have $m$ pivots, i.e. every row will have pivot and there will be $n-r=n-m$ free columns/variables. It’s typical reduced row-echelon matrix will look like $\begin{bmatrix} I & F \end{bmatrix}$In this case, there isn’t any restriction on $b$ as well as the equation will be solvable for each $b$ and will have infinitely many solutions.
-
For a case, when $r=m=n$, the matrix is a square matrix and we call the matrix a full rank matrix. A full rank rank matrix is always invertible. It should also be noted that it’s reduced row-echelon form is $I$. It’s nullspace is a zero vector and for the equation $Ax=b$ to be solvable, there is’t any restriction on $b$.
-
For a case when, $r < m;r < n$, the typical reduced row-echelon matrix will look like $\begin{bmatrix} I & F \\ 0 & F \end{bmatrix}$ and $Ax=b$ can have $0$ or infinitely many solutions.
One more thing to consider is the fact that for a $m \times n$ matrix $A$, if $m < n$, then we will have at least one non-zero solution to $Ax=0$.