The way to compute the inversion of a iteratively increased matrix in Matlab

Asked Nov 05 '17 at 22:03

Active Nov 28 '17 at 19:29

Viewed 57 times

For example, I have to compute the matrix A\B and the size of A and B is 3*3. After the first iteration, the size of the matrix become 4*4, however, the elements at (i,j) where i,j <= 3 are unchanged and the new data will be added to the location (m,n) where m,n=4. This process would repeat N times until reaching the maximum iterations. The most easy way is just computeA\B at each iteration in Matlab, but this is very inefficient. Since the new data will only add to the last row and column, and most of other elements in the matrix are not changed, I believe there is some other way to compute the results instead of re-computing the the whole A\B.

It would be so much appreciated that if someone could suggest and share some ideal to do this more effective.

edited Nov 28 '17 at 19:29

asked Nov 05 '17 at 22:03

Zhida Deng

2

Where do you iteratively increase `A` or `B`? Your code defines them from scratch at each iteration – Luis Mendo Nov 05 '17 at 23:28
Hi Luis, the code defines here only for describing my problem. In my problem, the matrix size will increase 1 at each iteration, and I need to re-computing the 'B\A'. – Zhida Deng Nov 05 '17 at 23:33
In your real problem, is `A` for `i = 1` a submatrix of `A` for `i = 2` ? – Leander Moesinger Nov 06 '17 at 09:27
Hi Leander, I have edited the code statement, which may be more clearly present my problem. Thank you! – Zhida Deng Nov 06 '17 at 10:50

The way to compute the inversion of a iteratively increased matrix in Matlab

0 Answers0