Genetic algorithm for optimizing service distributions

Abstract

This paper deals with optimizing the distribution of services performed by agents in a mobile agent network (MAN) using a genetic algorithm. The MAN is a model which consists of agents and various network elements and nodes. Each agent is capable of performing remote software management services. In order to verify the efficiency of various service distributions on different network topologies, we developed a mobile agent network simulator. This simulator allows the user to define the services which need to be performed by agents and describe the physical network used in the process. In this paper, an approach using a genetic algorithm is proposed. This algorithm is used to generate a suboptimal distribution of services among agents which optimizes the total execution time of services on a variety of network topologies with different network parameters.