A Novel Agent-Based Load Balancing Algorithm for Grid Computing

This paper presents an agent-based resource scheduling algorithm for Grid computing systems. With this scheduling algorithm, the computational resources of the Grid can be dynamically scheduled according to the real-time working load on each node. Thus a Grid system can hold an excellent load balance state. Furthermore, the application of this algorithm is introduced into the practical protein molecules docking applications, which run at the DDG, a Grid computing system for drug discovery and design. Solid experimental results show the load balance and robustness of the proposed algorithm.