摘要
Abstract
The task scheduling problem for heterogeneous multi-core processors has been proven to be NP-complete.To meet the computational demands of complex applications and enhance the efficiency of task scheduling in heterogeneous multi-core processors,an Evolutionary Adaptive Bat Algorithm(EABA)-based task scheduling algorithm for heterogeneous multi-core processors is proposed.First,the task scheduling problem is described,and a corresponding mathematical model is established.Subsequently,a task allocation encoding scheme and a fitness function are designed to map the proposed algorithm into a discrete space,making it suitable for studying discrete task scheduling problems in heterogeneous multi-core processors.Subsequently,to prevent the algorithm from prematurely converging to the local optima,a decaying pulse strategy and an evolutionary adaptive transformation strategy are introduced.Finally,simulation experiments are designed to compare the proposed algorithm with the Bat Algorithm(BA),Improved Particle Swarm Optimization(IPSO)algorithm,Artificial Fish Swarm Algorithm(AFSA),and Improved Whale Optimization Algorithm(IWOA).The experimental results demonstrate that,under medium-scale tasks(40 to 70 tasks)and large-scale tasks(80 to 100 tasks),the optimal scheduling length of the EABA is shortened by 12.86％and 13.67％,respectively,compared with that of the suboptimal algorithm,with average execution time reductions of 14.51％and 13.50％,respectively.关键词
异构多核/任务调度/蝙蝠算法/进化自适应/有向无环图Key words
heterogeneous multi-core/task scheduling/Bat Algorithm(BA)/evolutionary adaptive/Directed Acyclic Graph(DAG)分类
计算机与自动化
