From: lg@kt.dtu.dk (Lars Gregersen)
Subject: Re: Inverse of a matrix
Date: Mon, 13 Sep 1999 07:18:54 GMT
Newsgroups: sci.math.num-analysis
Keywords: complexity of matrix inversion

On 12 Sep 1999 15:54:19 +0200, Robert Harley <harley@corton.inria.fr>
wrote:

>
>wei-choon ng <wng@ux4.cso.uiuc.edu> writes:
>> 	A friend of mine told me that the fastest method of computing the
>> inverse of a matrix is of O(N^2.376).  He seemingly remembers one of the
>> guys who achieved this as Copper-something.  Does anyone have the
>> reference to this technique?
>
>Don Coppersmith and Shmuel Winograd.
>
>The method, for matrix multiplication, is described starting at page
>23 of in "Computational Algebraic Complexity", edited by Erich Kaltofen.
>
>You can find the relationship between multiplying and inverting
>matrices in Chapter 6 of Aho, Hopcroft and Ullman.

The inversion is treated in "Matrix computations" by Golub and van
Loan.

Have fun.

  Lars

------------------------------
Lars Gregersen (lg@kt.dtu.dk)
http://www.gbar.dtu.dk/~matlg