Gravitational Emulation Local Search Algorithm for Advanced Reservation and Scheduling in Grid Computing Systems

Abstract

The issue of the guarantee quality of service (QOS) for users can be provided by the advanced reservation. The advanced reservation is a kind of mechanism that can provide the ability to allocate resources for users based on the agreement upon the needs of quality service and increase of the number of accepted users' requests in Grid system. Scheduling and advanced reservation of the resources in grid system is a NP-complete issue, so the deterministic algorithms can not be used to improve it. Some heuristic methods for this purpose are: Genetic Algorithm (GA), Simulated Annealing (SA), Hill Climbing (HC), etc. In this paper, a heuristic method, called the Gravitational Emulation Local Search algorithm or GELS will be shown in solving of the scheduling and the advance reservation of the resources. The offered algorithm is named Gravitational Emulation Local Search Advanced Reservation algorithm or GELSAR and so as to confirm it, the algorithm should be implemented and compared with GA. The results show that the accepted jobs by applying GELSAR is increased 7.5 percent and the computation time of our algorithm can be lowered down to 50 percent compared to GA.