An efficient local search for grid scheduling problem in learning system

Grid scheduling is an NP-complete problem which the objective is to minimize the completion time (makespan). This paper is motivated by the requirement that the feedback analysis service of e-learning system needs to distribute the computing jobs for balancing the workload and getting the fast response. We propose an efficient local search method embed into the differential evolution algorithm (DE) to find a feasible solution. DE is efficiently in the computing time and the proposed local search method tries to enhance the result of DE for minimizing the makespan. Comparing with DE, our proposed algorithm reduces the makespan with the ratio 64% in the best case. Simulation results demonstrate the proposed method.