Research On Attribute Reduction And Partitional Clustering Based On Membrane Computing

Membrane computing is a new branch of natural computing.It is a research field inspired by the characteristics and functions of biological cell membranes to abstract computing models and methods.The research directions include theory research of membrane computing model,application research of membrane computing model,membrane computing model implementation,and membrane algorithm research.Since the rule execution in the membrane computing model has maximum parallelism,so that it can solve the NP-hard problem in polynomial time,the membrane computing model has been applied to many fields such as automatic control,economics,computer graphics and so on.Attribute reduction is a basic problem in data analysis and processing.Its purpose is to remove redundant and irrelevant attributes from the attribute set of the original data to preserve the optimal subset of attributes.Attribute reduction based on rough set theory is an important attribute reduction method.In recent years,a series of research results have been obtained.However,studies on attribute reduction based on membrane computing models have rarely been reported.In this paper,based on the cell-like P system model,an attribute reduction P system based on rough set theory is designed,which can solve the optimal attribute subset in polynomial time.Based on the evolutionary mechanism of biological cell membranes in its life cycle,this paper proposes a new membrane evolution algorithm framework MEAF?Membrane Evolutionary Framework?.Compared with traditional membrane algorithms,MEAF only needs to rely on its own evolutionary operator for calculation,while traditional membrane algorithms needs to be combined with other algorithms.We apply MEAF to the clustering problem and propose a new partitional clustering algorithm MECA?Membrane Evolutionary Clustering Algorithm?.The experimental results show that MECA has good performance.The main research work completed in this paper includes:?1?For the attribute reduction problem in data analysis and processing,a property reduction type cell P system?based on rough set theory is designed.The complexity analysis shows that it can solve all the optimal attribute subsets of the attribute reduction problem in polynomial time.The example analysis and simulation results verify the feasibility and effectiveness of?.?2?For the clustering problem,this paper proposes a new membrane-based evolutionary clustering algorithm MECA.Compared with the traditional membrane algorithm,it has its own evolutionary operator.MECA is used to analyze and compare other evolutionary algorithms.The experiment results verified the validity and stability of the MECA.This paper combines membrane computing with attribute reduction and cluster analysis to carry out related research work.The research results enrich the application field of membrane computing model.At the same time,the proposed partitioning clustering algorithm has certain reference value and significance for the research of evolutionary algorithm and clustering algorithm.