An Enhanced Attribute Co-evolutionary Game Reduction Algorithm by Integrating Self-adaptive Multi-level Nash Equilibrium  

作　　者：DING Weiping GUAN Zhijin SHI Quan 

机构地区：[1]School of Computer Science and Technology, Nantong University [2]State Key Laboratory for Novel Software Technology, Nanjing University

出　　处：《Chinese Journal of Electronics》2016年第1期13-19,共7页电子学报（英文版）

基　　金：supported by the National Natural Science Foundation of China(No.61300167,No.61171132);the Open Project Program of State Key Laboratory for Novel Software Technology,Nanjing University(No.KFKT2015B17);the Natural Science Foundation of Jiangsu Province(No.BK20151274);the Natural Science Pre-research Foundation of Nantong University(No.12ZY016);the Qing Lan Project of Jiangsu Province

摘　　要：In order to further analyze the dynamical behavior of co-evolutionary populations in minimum attribute reduction, an Enhanced attribute co-evolutionary game reduction(EACGR) algorithm by integrating selfadaptive multi-level Nash equilibrium is proposed in this paper. First, a self-adaptive multi-level Nash game model with cross co-evolution is designed to provide the better solution for the dynamical symmetric co-evolution of multi-populations. Second, the profit matrix of elitist energy is constructed to explore the payoff mechanism of coevolutionary selection. And then a novel Nash equilibrium strategy is adopted to perform attribute co-evolutionary game reduction so that the admissible balance of attribute reduction can be well achieved. Experimental results indicate that EACGR has the higher performance of minimum attribute reduction, and the application into 3D brain MRI segmentation with promising results indicates its stronger robustness and practicability.In order to further analyze the dynamical behavior of co-evolutionary populations in minimum attribute reduction, an Enhanced attribute co-evolutionary game reduction（EACGR） algorithm by integrating selfadaptive multi-level Nash equilibrium is proposed in this paper. First, a self-adaptive multi-level Nash game model with cross co-evolution is designed to provide the better solution for the dynamical symmetric co-evolution of multi-populations. Second, the profit matrix of elitist energy is constructed to explore the payoff mechanism of coevolutionary selection. And then a novel Nash equilibrium strategy is adopted to perform attribute co-evolutionary game reduction so that the admissible balance of attribute reduction can be well achieved. Experimental results indicate that EACGR has the higher performance of minimum attribute reduction, and the application into 3D brain MRI segmentation with promising results indicates its stronger robustness and practicability.

关 键 词：Cross co-evolutionary game Profit ma-trix of elitist energy Multi-level Nash equilibrium Mini-mum attribute reduction. 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]