Using a GA adaptor in multi-applications

Abstract

The topic discussed in this paper is a Genetic Algorithm (GA) for multi-applications without any user involvement in the process of the problem design. This user-friendly GA can be used as part of an intelligent tool for problem solving. The main idea of the design is to build a GA based on an adaptor, which encodes problems into a standardized format. This standardized format is then presented to the GA. Using this method; we can separate the GA from any particular application and ensures that it becomes a reusable tool. An extension of this concept is that we can build a number of format adaptors so that it can tackle a wider range of problems. We tested the GA using a number of combinatorial optimisation problems, particularly the Traveling Salesman Problem (TSP). The results indicated that the adaptor helped to improve the overall performance of the GA. © Springer-Verlag Berlin Heidelberg 2003.