Parallel genetic algorithms: an exploration of weather prediction through clustered computing
Emily Gibson, Jessie Burger · 2003
BACKGROUND In recent years, the price drop in off-the-shelf computer systems has enabled small institutions access to affordable supercomputing. Thus, there has been considerable growth in research into inherently parallel problems and methods, which are well suited to cluster computing environments. One of the most promising of these methods is the development of parallel genetic algorithms. Genetic algorithms are based on the theory of Darwinian evolution [Holland]. A large number of potential solutions to the problem are created and then competed against each other. The fittest solutions are combined to form new solutions and mutated to seek better solutions. Over time, an effective solution is developed [Xobitko]. Genetic algorithms are especially good with NP-complete and NP-hard problems, as well as problems which are not fully understood by the programmer. PROBLEM We are endeavoring to predict weather temperatures using genetic algorithms as a