Guided Ant Colony Optimization Based Variable Neighborhood...

GAVNS: Guided Ant Colony Optimization based Variable Neighborhood Search for Optimistic Load Balancing in Grid Computing Gurveer Kaur Brar1, Amit Chhabra2 1 Guru Nanak Dev University, Amritsar Punjab, India gurveer.dhillon43@gmail.com, amit.cse@gndu.ac.in Abstract Grid Computing resolves high performance computing and throughput issues through sharing of resources. These resources are heterogeneous in nature and geographically distributed to develop large scale applications. Scheduling and Load balancing is one of the key ideas in grid environment. For efficient scheduling, proper management of resources is required. This paper mainly presents Guided Ant Colony Optimization (ACO) based Variable Neighborhood Search (GAVNS) algorithm which represents how to schedule an independent jobs on grid nodes in order to optimize the schedule and load on nodes/servers. The performance of proposed algorithm is compared with existing Variable Neighborhood Search (VNS) algorithm. Simulation results have shown that GAVNS performs better than VNS. Moreover, better makespan is achieved through this technique. Keywords: Grid Computing; Job Scheduling; Makespan; Ant Colony Optimization; Load balancing; Variable Neighborhood Search 1. Introduction Grid Computing is a form of parallel and distributed computing that permits sharing, selection and collection of widely dispersed dynamic resources at run time. They mainly rely on their availability, performance, capability, cost and