BFKL, MM, Alpert-Rokhlin's transform, FFTW, ARPACK and all that

Abstract

The evolution equation for $q \bar q$ production introduced by Marchesini and Mueller posed some intriguing mathematical puzzles, both numerical and analytic. I give a detailed account of the numerical approach which led eventually to the exact solution. While part of the work was in fact along a wrong track, it turns out that some of the techniques involved are interesting in their own and applicable to many other problems, i.e. to the numerical study of Ricci flows.