The Multi-objective Cloud Tasks Scheduling Based on Hybrid Particle Swarm Optimization

2020 
This paper proposes a hybrid particle swarm optimization algorithm, namely HPSO, to reduce application’s total execution time (ie., makespan) and balance the computation load, by solving the multi-objective task scheduling problem in cloud environments. The proposed algorithm starts with the strategy of pre-generating search particles that can improve particles’ position quality during the search procedure by refining the solution space. In addition, we incorporate the strategy of artificial fish swarm algorithm into the position updating process of particles swarm optimization (PSO), providing various updating schemes to solve the multi-objective scheduling problem. Moreover, we define the weighted sum of the standardized total execution time and load balancing as the optimization goal of the scheduling problem. Experimental results demonstrate that compared with PSO and immune algorithm, our HPSO algorithm can achieve better scheduling result in terms of shorter makespan and smaller standard deviation of computation times among all computing resources.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []