About LU factorization

[salvomic]

(06-04-2015 09:18 AM)Paul Dale Wrote:  Have you tried reversing the decomposition to see if you get the original matrix back?

Looks like the prime swapped the second and third rows during the factorisation which changes the later results.

Earlier HP calculators use the Doolittle decomposition with partial pivots, I'd be surprised if the prime didn't do something similar.

- Pauli

hi Pauli,
swapping 2nd and 3rd row LU give same L and U and (obviously) [1,2,3] for the permutation matrix...

Prime has pivot() but this function returns only a column "pivoted", i.e.
pivot(,[[2,1,1], [4,-6,0], [-2,7,2]],1,1) returns
[[2,1,1], [0,-16,-4], [0,16,6,28]]
so then we must do a loop: divide by the nth pivot, take the minor (delete 1st row, 1st column), use pivot() again...
(In another post about pivots I'm trying to do a program, but I've still troubles)...