摘要
本文主要基于现代启发式差分算法讨论多处理机调度,多处理机调度是NP组合优化问题,目前多采用启发算法。差分进化算法是最近提出的进化算法,主要根据父代个体之间矢量差构造下一代,是一种全局优化搜索方式。本文考虑采用差分进化矢量优先级模型描述调度顺序进行调度,与模拟退火算法比较得到较好调度结果。
This paper presents a differential evolution algorithm to solve the multiprocessor scheduling problem with arbitrary task graph. Multiprocessor scheduling is a NP-hard problem, many papers employ heuristic. Differentia evolution is a novel heuristic approach for minimizing possibly nonlinear and non-differentiable continuous space functions. In this paper we introduce scheduling model using priority-based solution representation, which enables the differential evolution algorithm to be applied to scheduling problems. The results for arbitrary task graph are compared to simulate annealing algorithm.
出处
《微计算机信息》
北大核心
2006年第10S期184-186,共3页
Control & Automation
基金
河南省科技攻关项目(0424300007)资助
关键词
差分进化
模拟退火
多处理机调度
优先级
Differential evolution, simulated annealing, multiprocessor scheduling, priority
