HATMOG: an enhanced hybrid task assignment algorithm based on AHP-TOPSIS and multi-objective genetic in cloud computing.

In recent years, despite the rapid growth of cloud computing platforms, the technology confronts significant challenges including virtualization, load balancing, fault tolerance and most important of all, task scheduling. Considering the last challenge, because of high number of users and significant growth of number of tasks, there are some limitations such as MakeSpan, high resource utilization rate and executive costs in task scheduling algorithms. In this study, an effective method called HATMOG based on the smart hybrid of multiple criteria decision making algorithms of AHP-TOPSIS and Non-Dominated Sorting Genetic Algorithm (NSGAII) has been used to improve task scheduling in the cloud. The proposed method was done in two phases. In the first phase, the tasks which entered the cloud environment were placed in separate queues and then, according to task length, number of required processor elements and task expire time were sorted by AHP-TOPSIS algorithm in their priority queues. This phase helped on time assignment of more important tasks to the most appropriate virtual machines significantly that resulted in response time decrease and optimized resource utilization. In the second phase, the sorted tasks with prioritized queues were assigned to the appropriate virtual machines using NSGAII. Task assignment to virtual machine is an NP-Hard issue and NSGAII helped the efficiency improvement of cloud computing environment significantly because of the high convergence speed in finding close to optimal solution. The results of numerous simulations in Cloudism showed that the proposed method improved MakeSpan comparing TOPSIS-PSO, AHP-TOPSIS-PSO, NSGAII and PSO by 17.76, 155.73, 5.05 and 171.35 percent respectively and the average resource utilization by 15.94, 176.59, 4.83 and 176.65 percent respectively. [ABSTRACT FROM AUTHOR]