SamAnt:A Self Organization Approach for Dynamic Load Balancing in Grid

Abstract

Grid computing is likely to be the future of the global computer village. As a part of grid resource management load balancing has vital e ect on overall grid performance. To nd an e cient load balancing scheme is rather chal- lenging with the dynamic, distributed and vast nature of a grid. Ant Colony Optimization(ACO) has a good potential to address the prob- lem with its distributed and intrinsically parallel nature. And furthermore we believe that by hybrid the strong factors in these natural optimization techniques as ACO and Particle Swarm Optimization(PSO) would produce more powerful and e cient solutions. In this paper an ACO/PSO hybrid load balancing algorithm (samAnt) is proposed and it is implemented on a framework. Evaluation results show that it is quite worthwhile with its scalability, adaptability, simplicity and less overheads.