一种基于最大熵的构件库动态存储方法

Dynamic Storage of Component Lib Based on a Maximum Entropy Principle

本文依照ALOAF构件库框架参考模型，首先给出构件库的形式化描述；然后，根据构件库框架的元模型以及相应的数据模型，确定构件在构件库中的存储方式；进而，提出了一种基于最大熵的构件库的动态存储方法，概括为分类、模糊聚类、划分和排序。此方法改变了传统构件库的静态组织形式．基于最大熵的构件块划分、基于贝叶斯方法的构件动态排列和淘汰机制使构件库的存储更加合理。

According to the reference model of ALOAF, the formalizing description of component lib is presented in this paper. The component storage method can be given by means of the meta-model and data model of component lib framework. Based on the maximum Entropy and Bayesian, the dynamic storage method of component lib is put forward, which has four stages： classifying, fuzzy clustering, compartmentalizing and sorting.