In cloud computing, efficient multi-objective task scheduling, aiming at minimizing makespan, energy consumption,and load variance,remains a critical challenge due to the non-deterministic polynomial( NP)-completeness...In cloud computing, efficient multi-objective task scheduling, aiming at minimizing makespan, energy consumption,and load variance,remains a critical challenge due to the non-deterministic polynomial( NP)-completeness of the problem and the limitations of traditional algorithms like premature convergence. In this paper,a multi-strategy improved sparrow search algorithm( MISSA) was proposed to address these issues. MISSA integrates specular reflection learning for initial population optimization,nonlinear adaptive decay weights to balance global exploration and local exploitation,and an innovative strategy based on T-distribution mutation to enhance population diversity. Experimental results on benchmark functions and real cloud task scheduling scenarios using CloudSim demonstrate that MISSA outperforms comparative algorithms such as sparrow search algorithm( SSA),boosted sparrow search algorithm( BSSA),and genetic algorithm-grey wolf optimizer( GA-GWO),achieving significant reductions in makespan,energy consumption,and load variance. MISSA provides an effective solution for intelligent resource allocation in heterogeneous cloud environments,showcasing robust performance in complex multi-objective optimization tasks.展开更多
基金supported by the Key Research and Development Project of Heilongjiang Province (JD2023SJ20)。
文摘In cloud computing, efficient multi-objective task scheduling, aiming at minimizing makespan, energy consumption,and load variance,remains a critical challenge due to the non-deterministic polynomial( NP)-completeness of the problem and the limitations of traditional algorithms like premature convergence. In this paper,a multi-strategy improved sparrow search algorithm( MISSA) was proposed to address these issues. MISSA integrates specular reflection learning for initial population optimization,nonlinear adaptive decay weights to balance global exploration and local exploitation,and an innovative strategy based on T-distribution mutation to enhance population diversity. Experimental results on benchmark functions and real cloud task scheduling scenarios using CloudSim demonstrate that MISSA outperforms comparative algorithms such as sparrow search algorithm( SSA),boosted sparrow search algorithm( BSSA),and genetic algorithm-grey wolf optimizer( GA-GWO),achieving significant reductions in makespan,energy consumption,and load variance. MISSA provides an effective solution for intelligent resource allocation in heterogeneous cloud environments,showcasing robust performance in complex multi-objective optimization tasks.
