摘要
一般粗粒度并行遗传算法(CGGA)的性能受诸多因素的影响表现不尽如人意。以降低通信代价为主要目标,受物种金字塔模型的启发,设计了一种双阈值限制下的自调整堆结构,并对其堆调整具体操作进行了改进,以期望改进后算法中种群间的通信代价大幅度降低,优化收敛速度,提高算法效率。通过对遗传算法的几个典型测试函数通信量的分析和实验表明,基于该模型的并行遗传算法在降低通信代价、提高收敛速度、优化最终解方面收效明显。
Common coarse - grained genetic algorithrn（CGGA） has been criticized for many reasons. In this paper, focus on communication costs,gain the idea from creature specices pyramid structure and suggest a heap model limited under two valves expect to significantly reduce the communication costs between two groups. The expectation of migration costs and experiment on typical GA test functions in the last part of this essay all verify that this model could greatly decrease the cost of communication and accelerate the convergence speed.
出处
《计算机技术与发展》
2007年第10期105-108,112,共5页
Computer Technology and Development
基金
国家自然科学基金项目(60273043)
安徽省自然科学基金(050420204)
安徽省高校拔尖人才基金(05025102)
安徽省教育厅自然科学研究项目(2006KJ098B)
关键词
并行遗传算法
CGGA
通信代价
堆模型
parallel genetic algorithm
CGGA
communication cost
heap model
