摘要
基因表达式编程(GEP)算法是遗传计算家族中的新成员,它参考生物进化的基因遗传规律而提出。笔者先后介绍了GEP 的应用背景和基本理论,重点研究了在改善种群多样性问题的优化策略,设计了将 GEP 与小生境技术结合的关键技术和算法流程,用定量方法确定个体之间的差异情况并动态调整适应度值,使得算法能够有效解决早熟收敛难题。实验部分用复杂函数建模验证了改进方法在进化速度和模型拟合度等方面的优势,为现实工作提供了有价值的预测模型和指导数据。
Gene expression programming is a new member of genetic computing family and it is proposed by referring to gene inheritance rules. This paper introduces the application background and basis theories of GEP and lays its emphasis on the optimization strategy of improving population diversity. It designs the critical technology and algorithm flow for the combination of GEP and niche technique which can solve the difficulty in premature convergence through determining the individual difference and further adjusting the fitness value dynamically. In the experiment field,the new method,through complicated function modeling,is proved to be advantageous in evolutionary rate and model fitting,thus provide valuable predicting model and guidance data for practical work.
出处
《山东师范大学学报(自然科学版)》
CAS
2015年第3期58-62,共5页
Journal of Shandong Normal University(Natural Science)
基金
山东省科学技术发展计划项目(2012G0022207)
山东省科技厅星火计划项目(2013XH1703).
关键词
基因表达式编程
多样性
小生境
函数建模
gene expression programming
diversity
niche
function modeling