摘要
本文将用于求解组合优化问题的模拟退火法引入聚类分析、属性关系图同态、分段曲线拟合和特征选择等模式识别问题。(1)提出了一类新的聚类分析算法——模拟退火聚类法;(2)给出了一种模拟退火图同态的方案和实现算法——ALISOM;(3)详细地讨论了如何应用模拟退火组合优化法进行分段曲线拟合和特征选择.
Simulated annealing technique for solving combinatorial optimization problems has been applied to cluster analysis, isomorphisms of attributed relational graphs, piecewise curve fitting and feature selection. (1). A class of new clustering algorithms by simulated annealing are presented. (2). The problem of isomorphisms of attributed relational graph is treated by annealing simulation. An annealing isomorphism algorithm ALISOM is presented. (3). The applications of simulated annealing technique to piecewise curve fitting and feature selection are discussed in detail.
出处
《自动化学报》
EI
CSCD
北大核心
1989年第2期114-121,共8页
Acta Automatica Sinica
关键词
模式识别
模拟退火法
聚类分析
Simulated annealing
pattern recognition
cluster analysis
graph isomorphism
feature selection.
