作者: Xingguang Peng , Fubin Zhang , Demin Xu
DOI:
关键词: Machine learning 、 Process (engineering) 、 Identification (information) 、 Artificial intelligence 、 Population diversity 、 Compensation (engineering) 、 Compensation methods 、 Computer science 、 Evolutionary computation 、 Diversity (business)
摘要: This paper focuses on the effect of population diversity to environment identification-based memory scheme (EI-MMS) which heuristically compensates through storage and retrieving process historic information. We introduced several compensation measures combined them with EI-MMS based univariate marginal distribution algorithm (UMDA) from two aspects. First, a basic measure was used fight against inherent loss UMDA. Second, environment-triggered were added in sense dynamic environment. Based experimental results three test problems, dynamics corresponding UMDAs analyzed conclusions about how does affect performance environments drawn.