Dynamic Resource Allocation Using Improved Firefly Optimization Algorithm in Cloud Environment

Abstract

Today, cloud computing has provided a suitable platform to meet the computing needs of users. One of the most important challenges facing cloud computing is Dynamic Resource Allocation (DSA), which is in the NP-Hard class. One of the goals of the DSA is to utilization resources efficiently and maximize productivity. In this paper, an improved Firefly algorithm based on load balancing optimization is introduced to solve the DSA problem called IFA-DSA. In addition to balancing workloads between existing virtual machines, IFA-DSA also reduces completion time by selecting appropriate objectives in the fitness function. The best sequence of tasks for resource allocation is formulated as a multi-objective problem. The intended objectives are load balancing, completion time, average runtime, and migration rate. In order to improve the initial population creation in the firefly algorithm, a heuristic method is used instead of a random approach. In the heuristic method, the initial population is created based on the priority of tasks, where the priority of each task is determined based on the pay as you use model and a fuzzy approach. The results of the experiments show the superiority of the proposed method in the makespan criterion over the ICFA method by an average of 3%.