Sunday, 5 December 2010

Numerical methods: matrix inversion

The F#.NET Journal just published an article about numerical methods:

"Matrix inversion is one of the most important numerical methods with many practical applications. In practice, the numerical instability of matrix inverses leads to the use of linear solvers that effectively premultiply by the inverse of a matrix without having to explicitly compute the matrix inverse. This article disregards numerical accuracy in order to write the simplest possible function that computes the inverse of a matrix. The resulting function is just 8 lines of F# code and the remainder of the article is devoted to examining the generality and parallelization of this tiny function. We find that our tiny solution is up to 4.3× faster than Mathematica 7's built-in matrix inversion..."

To read this article and more, subscribe to The F#.NET Journal today!

1 comment:

Art said...

Thanks Jon. Another valuable F#J article.

If you like this F#J article you may also enjoy Paul Jackson's Love The Dot blog ---
http://www.lovethedot.net/2010/02/parallel-programming-in-net-40-and.html