General applicability of genetic and simulated annealing algorithms for data mapping

Abstract

We experimentally analyze the general applicability of genetic algorithms (GA) and simulated annealing algorithms (SA) for mapping data to multicomputers. The results show that the GA and SA are insensitive to user parameters in wide ranges, completely fault tolerant, and unbiased towards particular multicomputer topologies. These properties of flexibility and general applicability, which are lacking in other heuristic algorithms, make the GA and SA attractive for automatic parallelization systems.