When we reduce this $Ax=b$ system repres. by an augmented matrix to row reduced echelon form, why do all components of the $b$ vector become numbers?

121 Views Asked by At

Consider the linear system $Ax=b$ where

$$A=\begin{bmatrix} 1&2&3&5\\ 2&4&8&12\\ 3&6&7&13 \end{bmatrix}\tag{1}$$

$$b=\begin{bmatrix} b_1\\ b_2\\ b_3 \end{bmatrix}\tag{2}$$

Let's use the augmented matrix

$$M=\begin{bmatrix} 1&2&3&5&b_1\\ 2&4&8&12&b_2\\ 3&6&7&13&b_3 \end{bmatrix}\tag{3}$$

to perform row operations.

I did the calculations that follow in the program Maple.

If we perform Gaussian elimination we obtain the matrix

$$U=\begin{bmatrix} 1&2&3&5&b_1\\ 0&0&2&2&b_2-2b_1\\ 0&0&0&0&b_3-5b_1+b_2\end{bmatrix}\tag{4}$$

My question regards what the reduced row echelon form of this matrix is. Maple tells me it is

$$R=\begin{bmatrix} 1&2&0&2&0\\ 0&0&1&1&0\\ 0&0&0&0&1\end{bmatrix}\tag{5}$$

I can't quite understand the last column. When I do the calculations by hand I get the following as the last column (starting from (4), divide the second row by $2$, then subtract $3$ times the second row from the first row)

$$\begin{bmatrix} b_1-3\left (\frac{b_2}{2}-b_1\right )\\ \frac{b_2}{2}-b_1\\ b_3-5b_1+b_2\end{bmatrix}\tag{6}$$

What am I missing?

1

There are 1 best solutions below

0
On

User @jgd1729 provided the answer in the comments, and I want to add some information here about what I did in Maple to obtain the expected results.

I was using the command

LinearAlgebra:-ReducedRowEchelonForm(M)

but this reduced the entire augmented matrix including the last column representing $b$.

To obtain the $b$ that would be associated with the reduced row echelon system I did the following

P, L, U1, R := LinearAlgebra:-LUDecomposition(A, method = 'RREF')

This command provides the matrices in the factorization

$$A=P\cdot L\cdot U1\cdot R$$

where $P$ is a permutation matrix representing row exchanges, $L$ is the lower triangular inverse of the elimination matrix used to perform Gaussian elimination, $U1\cdot R$ is the Gaussian elimination reduced matrix, and $R$ by itself is the reduced row echelon form of $A$.

We can write

$$R=(P\cdot L\cdot U1)^{-1}A$$

Thus, when we go from $Ax=b$ to $Rx=d$, $d$ is

$$d=(P\cdot L\cdot U1)^{-1}b$$

In Maple, this is

d := (P . L . U1)^(-1) . b

which gives the result I got by hand

enter image description here