Easiest way to solve lower hessenberg matrix with Gaussian Elimination?

237 Views Asked by At

This question has been asked before, but in this case, we should do a little bit different thing.

Assume that we have this lower Hessenberg matrix.

$$A = \begin{bmatrix} CB & 0 &0 & 0 & 0\\ CAB & CB & 0& 0 &0 \\ CA^2B &CAB & CB & 0 & 0\\ CA^3B & CA^2B &CAB & CB & 0\\ CA^4B & CA^3B & CA^2B & CAB & CB \end{bmatrix}$$

Where $CB \in \Re^{n x n}$ and $CA^iB \in \Re^{n x n}$ and $n = 1$

We want to solve $Ax = b$, classic. But what if we have $n > 1$ ? E.g matrix $A$ looks like this:

     0.07741     0.23210     0.00000     0.00000     0.00000     0.00000     0.00000     0.00000     0.00000     0.00000
     0.23865   123.33000     0.00000     0.00000     0.00000     0.00000     0.00000     0.00000     0.00000     0.00000
     0.12238    29.62902     0.07741     0.23210     0.00000     0.00000     0.00000     0.00000     0.00000     0.00000
    -0.00611    15.82693     0.23865   123.33000     0.00000     0.00000     0.00000     0.00000     0.00000     0.00000
     0.10198    28.81903     0.12238    29.62902     0.07741     0.23210     0.00000     0.00000     0.00000     0.00000
    -0.05920   -12.09672    -0.00611    15.82693     0.23865   123.33000     0.00000     0.00000     0.00000     0.00000
     0.07206    21.47042     0.10198    28.81903     0.12238    29.62902     0.07741     0.23210     0.00000     0.00000
    -0.05632   -15.31863    -0.05920   -12.09672    -0.00611    15.82693     0.23865   123.33000     0.00000     0.00000
     0.04746    14.49060     0.07206    21.47042     0.10198    28.81903     0.12238    29.62902     0.07741     0.23210
    -0.04167   -12.22749    -0.05632   -15.31863    -0.05920   -12.09672    -0.00611    15.82693     0.23865   123.33000

Here we can see that the dimension and column dimension is $n=2$. Solving if $n=1$ is easy. Then I use Gaussian Elimination.

But how should I do in this case when $n > 1$ and $A$ is still lower traingular hessenbeg matrix? What is the easiest way in this case when $n > 1$?

1

There are 1 best solutions below

12
On BEST ANSWER

If you have a system like this you can use forward substitution, $Ax=b$ then

$$x_{1} = \frac{b_{1}}{a_{1,1}} $$

and

$$ x_{i} = \frac{b_{i} - \sum_{j=1}^{i-1} a_{i,j} x_{j} }{ a_{i,i} }$$