Task Scheduling for Computational Grids Using NSGA II with Fuzzy Variance Based Crossover
Reza Salimi, Navid Bazrkar, Mostafa Nemati · Advances in Computing · 2013
Scheduling algorith ms have essential ro le in co mputational grids for managing jobs, and assigning them to appropriate resources. An efficient task scheduling algorith m can reduce the total Time and Price for jobs execution and improve the Load balancing between resources in the grid. In this paper, we address scheduling problem of independent tasks in the market-based grid environment. We use NSGA-II to optimize task scheduling problem in grid. For decreasing computation, we considered Load balancing problem and improved it in task scheduling indirectly using fu zzy system without imp lementing third objective function. For the first time, we proposed Variance based Fuzzy Crossover operator for this purpose and more variety in Pareto-optimal solutions. Two functions are defined to generate two inputs for fuzzy system. Variance of Costs and presence of resources in scheduling are used to specify probability of crossover intelligently. Second fuzzy function with cooperation of Makespan objective satisfies load balancing objective indirect ly. Our method conducts the algorith m toward best and most appropriate solutions with load balancing in less iteration. Results obtained proved that our innovative algorithm converges to Pareto-optimal solutions faster and with more quality.