Job online scheduling within dynamic grid environment

Abstract

This paper proposes the idea of adaptive job scheduling algorithm by using hybrid Ant Colony Optimization (ACO) and Tabu algorithms. The idea behind the scheduling algorithm is evaluation of completion time of jobs in a service Grid. The algorithm comprises of two main techniques; first of all, Grid Information Service (GIS) collects information from each grid node, ACO evaluates complete time of jobs in possible grid nodes and then assigns job to appropriate grid node. ACO is used to minimize the average completion time of jobs through optimal job allocation on each node as well. While, Tabu algorithm is used to adjust performance of grid system because online jobs are submitted to grid system from time to time. This paper shows that the algorithm can find an optimal processor for each machine to allocate to a job that minimizes the tardiness time of a job when the job is scheduled in the system.