Using Multi-objective Optimization Algorithm in Heterogeneous Grid Environment

Aggregating various resources of different organizations, Grid provides excess computation power and storage capacity and is a trend to solve the computation-intensive engineering and mathematical problem. Because of the dynamic and heterogeneous characteristic of Grid system, scheduling algorithms make a vital difference to the efficiency of Grid. In this paper, a heuristic algorithm based on Tabu Search is proposed to optimize multi objectives for Grid scheduling, minimizing the completion time and improving load balancing. The process of scheduling is divided into partial scheduling and every scheduling cycle is triggered by special event. Meanwhile, a self-adaptive neighbor search is used to deal with the stochastic resources and scheduling parameters are adjusted to fluctuant load in partial scheduling. The algorithm is simulated in GridSim toolkit and the results demonstrate better performance compared to other scheduling techniques.