项目名称: 基于新型小生境策略的多模态、多目标、动态进化算法的研究
项目编号: No.61305080
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 瞿博阳
作者单位: 中原工学院
项目金额: 25万元
中文摘要: 智能优化计算的研究与应用是近年来的热点, 尤其是针对多模态(multi-modal)、多目标(multi-objective)以及动态(dynamic)等复杂问题的优化算法。 本项目以小生境(niching)策略为基础,对优化种群的多样性以及并行优化的收敛性进行 研究与探索,从而分析并提高算法的搜索性能,为创建新型有效的多模态、多目标及动态智 能优化算法提供思路,并为这些算法在实际中的应用奠定基础。项目的研究从多模态智能优 化算法着手,用自适应邻域限制的方法,阐明小生境方法和种群多样性在多模态优化问题中 的重要性,并进一步揭示多模态优化问题与多目标优化问题及动态优化问题的相似处及重要 关联关系,从而构建高效稳定的多模态、多目标及动态智能优化算法,并将最终提出的算法 应用在电力调度、资产配置等实际问题中,为高效解决实际复杂优化问题提供理论依据和核 心技术。
中文关键词: 进化计算;多模态优化;多目标优化;约束优化;
英文摘要: Evolutionary computation becomes an active research area in recent years, especially on multi-modal, multi-objective and dynamic optimization. This project uses niching techniques to improve the population diversity and converging speed. The proposed niching technique is incorporated into multi-modal, multi-objective and dynamic optimization algorithms to improve their performances.The project starts with the research on novel niching techniques on multi-modal optimization. The niching techniques focus on self-adaptive neighborhood mutation method which reveals the importance of niching method in increasing the diversity of the population. The research further explains the relationships among multi-modal, multi-objective and dynamic optimizations and subsequently construct novel multi-modal, multi-objective, dynamic evolutionary optimization algorithms. Finally, the proposed algorithms will be applied on power dispatch, asset allocation, etc. problems and these applications provide theories and techniques for complex real world optimization.
英文关键词: Evolutionary computation;Multi-modal optimization;multi-objective optimization;Constrained optimization;