From: Hans D Mittelmann Newsgroups: sci.math.num-analysis,sci.op-research,ucam.cl.misc Subject: Re: Wanted: Simulated Annealing reference Date: Tue, 22 Oct 1996 07:49:17 -0700 Peter Sewell wrote: > > I have a real-world optimisation problem, to find good (not necessarily the > least) minima of a complicated function of approx 3 -- 30 bounded > real variables. > > Not having touched a real number in years, I'm looking for a good > pragmatic book or tutorial paper on simulated annealing, or more generally on > global optimisation. > > Can anybody point me towards the one true reference? (I alreadly > have the sci.op-research faq.) > > Please respond by email - I'll summarise the responses if asked. > > thanks-in-advance, > > Peter Sewell Hi, you say you don't want necessarily the least minima but then still mention SA and GO. In case you really don't need the global minima, have not more than 30 variables and complicated functions which you may not want to differentiate, which may not even be everywhere differentiable, then I could point to an algorithm developed right where you are. It is available at ftp://plato.la.asu.edu/pub/other_software/COBYLA.tar.gz (f77) or ftp://plato.la.asu.edu/pub/other_software/COBYLA_C.tar.gz (C) For your information: there are also deterministic methods for global optimization not only stochastic ones. A standard reference on the former is Horst&Tuy. -- Hans D. Mittelmann http://plato.la.asu.edu/ Arizona State University Phone: (602) 965-6595 Department of Mathematics Fax: (602) 965-0461 Tempe, AZ 85287-1804 email: mittelmann@asu.edu