Effect of Global Parallelism on the Behavior of a Steady State Genetic Algorithm for Design Optimization

In this paper we investigate the effect of global parallelism using a master slave approach, on the behavior of a steady state genetic algorithm for design optimization. Empirical results in several engineering design domains demonstrate that this simple form of parallelism which has the potential for almost linear speedup, does not significantly disturb the convergence pattern of the GA even when the number of processors is comparable to the size of the population.

The Extended Abstract is published in the Proceedings of the Evolutionary Computing and Parallel Processing Workshop at GECCO-99, Orlando, FL, July 13, 1999.

The Full Paper is published in the Proceedings of the Congress on Evolutionary Computation (CEC99), Washington, DC, July 6-9.

Back to Brian Davison's publications