Improvement of Metaheuristic Methods for Task Scheduling in Cloud Computing
Improvement of Metaheuristic Methods for Task Scheduling in Cloud Computing
azime nasiripour1 Dr. Fazlollah Adibnia2 Dr. AliMohammad Latif3
1) Information Technology Eng student, Computer Engineering Department , Yazd University, Yazd, Iran
2) Associate Professor, Computer Engineering Department - Information TechnologyYazd University, Yazd, Iran
3) Associate Professor, Computer Engineering Department - Artificial intelligence
IOI Code
:
XCCA-FFHHD
Publication :
9th International Conference on Advanced Research in Science, Engineering and Technology(setconf.ir)
Abstract :
Cloud computing allows users to use distributed computing resources as services on the network. In spite of limited and heterogeneous resources and varied user requests for both, cloud providers and users an appropriately organized scheduling are required. Optimal scheduling for mapping tasks to resources is a key issue in cloud computing. Criteria such as makespan, response time, waiting time, energy consumption, load balancing, productivity rate, and the other criteria are considered in different scheduling methods. The problem of task scheduling is an NP-complete problem, so finding an exact solution, especially in large scale, is not feasible. This paper proposes an algorithm that is the combination of genetic algorithms and the Discrete Symbiotic Organism Search Algorithm (DSOS), inspired by the symbiotic of living organisms. The proper features of the genetic algorithm and utilizing this algorithm in the right place for such a problem improve the functionality of the system. The results show due to finding better solutions, the proposed method improves the average makespan by 22%, response time by 24%, waiting time by 25%, and utilization rate by 22% compared to the DSOS algorithm.
Keywords : cloud computing task scheduling symbiotic organism search genetic