并行化遗传算法研究综述

REVIEW OF PARALLEL GENETIC ALGORITHM

说明遗传算法的基本思想和特点。根据近五年国内遗传算法的研究现状,分析遗传算法当前发展的潜力与不足。对遗传算法未来的发展和研究热点进行了推理,指出遗传算法的主要发展方向是并行化,研究热点将集中在早熟机理和参数设置等方面,并且遗传算法未来会跟其他的技术进一步结合。从遗传算法的主要环节入手,分析遗传算法的并行化策略和4种常见的并行化模型,并分析不同模型使用的硬件环境和模型的优缺点。对并行化遗传算法的评价模型进行讨论,说明了常见的评价模型和改进之处。

The basic idea and characteristics of genetic algorithm were described in this paper.According to the research status of genetic algorithm in China in recent five years,we analyzed the potential and deficiency of genetic algorithm,and reasoned the future development and research hotspot of genetic algorithm.It was pointed out that the main development direction of genetic algorithm was parallelization,and the research hotspot focused on the mechanism of precocity and parameter setting.The genetic algorithm would be further combined with other technologies in the future.Starting from the key steps of genetic algorithm,we analyzed the parallel strategies and four common parallel models,and analyzed the hardware environment and advantages and disadvantages of different models.The evaluation model of parallel genetic algorithm was discussed,and the common evaluation model and improvement were explained.