Recall that we modeled the descent path with the cubic polynomial
Solve for the vector .
\begin{proof}[Solution.]
To start, we need to transform matrix into 17.2 Row Echelon Form by multiplying on the left by elementary matrices. Notice that rows 1 and 3 are nice looking from the standpoint of elimination, while rows 2 and 4 look like they require more work for elimination. Let's switch row 3 to row 2, row 4 to row 3, and switch row 2 to row 4 using a permutation matrix (combination of 9.10 Transposition Matrix).
We now have a structure that will be much easier to transform into row echelon form. We can zero out the bottom left corner by multiplying on the left by a 9.8 Shear Matrix. Recall that we get by taking the negative of what we are trying to cancel out and dividing by its pivot.
Now, we need to cancel out the values below the diagonal in column 2. Once again, we multiply by a shear matrix.
Now, we need to eliminate the sub diagonal values in column 3. We see that we just need to eliminate 160000.
We have successfully transformed our original modeling matrix into row echelon form . Using 12.2 Backwards Substitution, we produce the following values of :
We can conclude that the descent path is modeled by the cubic polynomial