面向多核系统的蚁群最优化能耗调度算法

Ant- colony Optimization Energy- saving Scheduling Algorithm for Multi- core Systems

主要研究多核系统的能耗最优化问题。首先,根据多核系统的特点,建立关于多核系统的任务、能耗模型;接着,设计和实现一种基于蚁群最优化能耗调度算法。本文算法以蚁群算法的概率状态转移规则为核心,通过全局信息激素更新的策略,避免了陷入局部最优的情况,从而获得全局最优解。通过实验比对,分别与贪心算法、穷举算法比较后发现:本校提出的调度算法,在最佳情况下,仅比最优能耗高0.7%,而该算法复杂度低,可以应用于任务输入集较大的场景。

This paper mainly studies the energy consumption optimization on the multi- core systems. Firstly,according to the characteristics of the multi- core systems,we formulated the task model and energy consumption model on the multi- core systems. Then,based on ant- colony optimization,we designed and implemented an energy saving algorithm on multi- core systems. The core idea of the proposed algorithm is to use the transition rules of probabilistic state to obtain the proper solution. Furthermore,the global pheromones update strategy is used to avoid falling into local optimum. The experimental results show that the proposed algorithm outperforms better than that of greedy algorithm,etc. And it has only 0. 7% higher energy consumption than that of the exhaustive algorithm. Moreover,it has a low- complexity in the large- scale data set.