Agent Based Load Balancing in Grid Computing

Abstract

Load-Balancing is an important problem in distributed heterogeneous systems. In this paper, an Agent-based load-balancing model is developed for implementation in a grid environment. Load balancing is realized via migration of worker agents from overloaded resources to underloaded ones. The proposed model purposes to take benefit of the multi-agent system characteristics to create an autonomous system. The Agent-based load balancing model is implemented using JADE (Java Agent Development Framework) and Alea 2 as a grid simulator. The use of MAS is discussed, concerning the solutions adopted for gathering information policy, location policy, selection policy, worker agents migration, and load balancing.