Task scheduling in computational grid using nature iinspired optimization techniques

Abstract

Grid computing is a new distributed heterogeneous computing paradigm that aims at newlineachieving resource sharing and collaborative computing. The resources in the Grid are grouped newlinetogether to form a virtual organization that is applied to solve a large scientific or business newlineapplication. The primary objective of this technology is to share idle and geographically newlinedistributed resources such as computational power and storage capacity. newlineTask (or job) scheduling in computational Grid, that is, the allocation of user submitted newlinetasks to appropriate distributed computational resources, is one of the most challenging and newlinecomplex task. In other words, the task scheduling in computational Grid is considered as NP- newlinehard problem due to the problem complexity and intractable nature of the problem. Such a newlineproblem could be solved using meta-heuristic algorithms. These types of algorithms have the newlineability to find near optimal solution in reasonable time rather than optimal solution in a very long newlineprocessing time. newlineThe primary objective of the scheduling is to minimize the makespan of the system. newlineMakespan is an indicator of the general productivity of the Grid system: small values of newlinemakespan mean that the scheduler is providing good and efficient planning of tasks to resources. newlineConsidering makespan as a stand-alone criterion not necessarily implies optimization of other newlineobjectives. Other important objectives are minimizing flowtime, processing cost, and job failure newlinerate and maximizing resource utilization rate. Various research works have been done on task newlinescheduling problem in Grid, but still further analysis and research needs to be done to improve newlinethe performance of scheduling algorithm in computational Grid. Hence in this thesis, we have newlinespecifically focused on improving computational Grid performance in terms of makespan, newlineflowtime, processing cost, job failure rate and resource utilization rate. newline