基于基因表达式编程的智能模型库系统的实现

Implementation of Intelligent Model Base System Based on Genetic Expression Programming

模型库是决策支持系统(DSS)的重要组成部分。传统的智能模型库系统对先验知识依赖性很强,难以实现真正意义上的智能化。针对这一弱点,提出了显式智能模型,隐式智能模型,显式模型基因的概念,给出了基于GEP的隐式智能模型(ImplicitIntelligentModel,IIM)的挖掘算法(GEP-IIMA),在此基础上实现智能模型库系统;研究了隐式智能模型库(IIMB)系统与GIS的接口技术。IIMB是真正意义的无先验知识的智能模型库,其模型的类型和参数的求解均由程序自己来实现。通过比较实验,系统所挖掘的函数模型比传统方法所得模型的精度提高了14.9%。

Model Base is important to Decision Support System. Traditional intelligent model base system depends on transcendental knowledge, and it is difficult to authentically implement intelligentization. To solve the problem, this paper makes following contributions: 1) Proposes the concepts of Explicit Intelligent Model, Implicit Intelligent Model, and Explicit Gene; 2) Proposes Mining Implicit Intelligent Model Algorithm Based on Gene Expression Programming (GEP-IIMA) and implements the intelligent model base system based on GEP-IIMA; 3) Studies the interface between the Implicit Intelligent Model Base (IIMB) system and GIS; 4) Gives compared experiments. IIMB is indeed an intelligent model base on it both the types and parameters of model are determined automatically by the GEP-IIMA. The comparison experiments show that the precision of the function model found by the GEP-IIMA is improved 14.9% than traditional method.