Maximizing availability and minimizing markesan for task scheduling in grid computing using NSGA II

Abstract

Large distributed platform for computationally exhaustive applications is provided by the Computational Grid (CG). Required jobs are allotted to the computational grid nodes in grid scheduling in order to optimize few characteristic qualities of service parameters. Availability is the most important parameter of the computational nodes which is the likelihood of computational nodes accessible for service in specified period of time. In this paper, emphasis has given on optimization of two quality of service (QoS) parameter makespan (MS) and availability grid system for the task execution. Since, the scheduling problem is NP-Hard, so a meta-heuristics-based evolutionary techniques are often applied to solve this. We have proposed NSGA II for this purpose. The performance estimation of the proposed Availability Aware NSGA II (AANSGA II) has been done by writing program in Java and integrated with gridsim. The simulation results evaluate the performance of the proposed algorithm.