Optimizing Multiple QoS for Workflow Applications using PSO and Min-Max Strategy

International Research and Innovation Summit 2017

Ambursa, Faruku Umar^1 ; Latip, Rohaya^1,2 ; Abdullah, Azizol^1 ; Subramaniam, Shamala^1

Communication Technology and Network Department, Faculty of Computer Science and Information Technology, Universiti Putra Malaysia, Malaysia^1

Institute for Mathematical Research (INSPEM), Universiti Putra Malaysia, Malaysia^2

关键词: Improve performance;    Meta-heuristic techniques;    New particle swarm optimization;    Optimization problems;    QoS requirements;    User requirements;    Workflow applications;    Workflow scheduling;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/226/1/012120/pdf
DOI  :  10.1088/1757-899X/226/1/012120

来源: IOP

【 摘 要 】

Workflow scheduling under multiple QoS constraints is a complicated optimization problem. Metaheuristic techniques are excellent approaches used in dealing with such problem. Many metaheuristic based algorithms have been proposed, that considers various economic and trustworthy QoS dimensions. However, most of these approaches lead to high violation of user-defined QoS requirements in tight situation. Recently, a new Particle Swarm Optimization (PSO)-based QoS-aware workflow scheduling strategy (LAPSO) is proposed to improve performance in such situations. LAPSO algorithm is designed based on synergy between a violation handling method and a hybrid of PSO and min-max heuristic. Simulation results showed a great potential of LAPSO algorithm to handling user requirements even in tight situations. In this paper, the performance of the algorithm is anlysed further. Specifically, the impact of the min-max strategy on the performance of the algorithm is revealed. This is achieved by removing the violation handling from the operation of the algorithm. The results show that LAPSO based on only the min-max method still outperforms the benchmark, even though the LAPSO with the violation handling performs more significantly better.