Structure and Performance of Fine-Grain Parallelism in Genetic Search

Within the parallel genetic algorithm framework, there currently exists a growing dichotomy between coarse-pain and fine-grain parallel architectures. This paper attempts to characterize the need for fine-grain parallelism. and to introduce and comparethree modelsof fine-grain parallel genetic algorithms (GAS). The performance of the three models is examined on seventeen test problems and is compared to the performance of a coarse-grain parailel GA. Preliminary results indicate that the massive distribution of the fine-grain parallel GA and the modified population topology yield improvements in speed and in the number of evaluations required to find global optima.

• Type: Conference
• Year: 1993
• Where: ICGA
• Authors: Shumeet Baluja
• Keywords: Fine-grain Parallel, Fine-grain Parallel Architectures, ICGA 1993, ICGA 1997, Parallel Genetic Algorithms