Multi Ant LA: An adaptive multi agent resource discovery for peer to peer grid systems

Grid system has become an important tool in solving large complex problems in recent years. This has led to further growth in the infrastructure. The centralized or hierarchical approach in the discovery of resources has failed to provide the required efficiency in the infrastructure. Although the synergy between grid and Peer to peer (P2P) systems was explored to address the problem encountered in the conventional resource discovery approaches, they are however still faced with issues ranging from network flooding to poor performance in dynamic networks. This paper proposes a mechanism based on the Learning Automata (LA) and Ant Colony Optimization (ACO) for resource discovery in a grid infrastructure. The Mobile Ants provided by the ACO locates the shortest paths within the grid system and the LA selects the optimal path for the mobile ants decision making. While compared to some existing resource discovery approaches, the proposed mechanism showed an improved performance.