GENERALIZED DISTRIBUTED GENETIC ALGORITHM FOR OPTIMIZATION PROBLEMS

The Genetic Algorithm has been used for optimization problems in many areas. One of the attractive features of the Genetic Algorithm is that it lends itself very well to parallel and distributed processing. Thi s feature of the Genetic Algorithm is used in this paper for improving its performance for large and complex optimization problems by implem enting it in a distributed environment. The key attribute of the distr ibuted implementation is that it can be used for different types of op timization problems without any modifications. In addition, the Distri buted Genetic Algorithm implementation provides fault tolerance by aut omatically redistributing the work load assigned to the failed process or(s). This redistribution of load is carried out in a user transparen t manner. The effectiveness and generality of the Distributed Genetic Algorithm implementation is demonstrated by solving several problems s uch as network topology design, network expansion and file allocation. (C) 1997 John Wiley & Sons, Inc.