A New Parallel Genetic Algorithm Model

Abstract

This paper presents an implementation of three Genetic Algorithm models for solving a reliability optimization problem for a redundancy system with several failure modes, a modification on a parallel a genetic algorithm model and a new parallel genetic algorithm model. These three models are: a sequential model, a modified global parallel genetic algorithm model and a new proposed parallel genetic algorithm model we called the Trigger Model (TM). The reduction of the implementation processing time is the basic motivation of genetic algorithms parallelization. In this work, parallel virtual machine (PVM), which is a portable message-passing programming system, designed to link separate host machines to form a virtual machine which is a single, manageable computing resource, is used in a distributed heterogeneous environment. The best result was reached and The TM model was clearly performing better than the other two models.