Enhanced Hyper-Heuristic Scheduling Algorithm for Cloud

Abstract

Task scheduling in cloud is mainly focused to find better and optimal solutions in order to minimize the total processing time of Virtual Machines. One of the objectives is to allocate availabe resources to tasks such that the execution of tasks is completed in minimal time with efficient use of resources. Task scheduling problem in cloud is known as NP-complete. One feasible solution for these type of problems is to apply hyper-heuristics. Hyper-heuristics are high level methods for solving complex problems that works on a search space of underlying heuristics. Previous studies have shown that using hyper-heuristic scheduling algorithm in cloud computing produces improved results. In this paper an intelligent selection operator and a multi point crossover operator are introduced. The intelligent selection operator uses a time weight to penalize slower heuristics, so that better heuristics are selected. The multipoint crossover operator is used to combine two solutions to get diversified and possibly improved new solutions. The proposed approach has been implemented in CloudSim and compared against the other standard algorithms. It is observed that for large number of tasks the proposed algorithm has performed 10.67% to 20.75% better than other standard algorithms.