HYBRID CAT SWARM OPTIMIZATION SCHEME FOR NON-PREEMPTIVE SCHEDULING OF INDEPENDENT TASK IN CLOUD COMPUTING

Abstract

The growing number of customers that are requesting computation-based-resources to meet the increasing demand of resource hungry applications have spark a greater challenge on how effective can scheduling can be carried out at the cloud datacenters. Recent advancement in the uses of metaheuristics techniques are promising approach in scheduling resources to hungry applications, but however, are limited in their performances due to issues like premature convergence. To overcome this concern with the aim to provide an effective scheduling, we propose a non-preemptive Hybrid Cat Swarm Optimization Scheme (HCSOS) to serve as an ideal solution. In the proposed scheme, orthogonal Taguchi approach is incorporated to overcome premature convergence, and minimizes local and global imbalance, while Pareto dominant strategy is used for providing customers with the option of selecting their service preferences. The results of the simulation on CloudSim tool show that our proposed scheme compared to the benchmarked schemes can achieve a minimum total execution time and cost (with up to 42.87%, 35.47%, 25.49% and 38.62%, 35.32%, 25.56% reduction). We further unveiled that a statistical analysis based on 95% confidence interval shows our proposed HCSOS scheme is remarkable in term of efficiency.