基于GA算法的智能导购系统的研究与实现  

The Study and Realization of Intelligent Guide System Base on Genetic Algorithm

作　　者：王婷婷[1] 王志良[1] 

机构地区：[1]北京科技大学信息工程学院,北京100083

出　　处：《计算机工程与应用》2006年第22期214-216,共3页Computer Engineering and Applications

基　　金：国家自然科学基金资助项目(编号:60573059);北京市"现代信息科学与网络技术"重点实验室资助项目(编号:TDXX0503);北京科技大学重点基金资助项目

摘　　要：论文以服装导购系统为例,以模拟顾客思维方式、喜好特点为目标,搭建了一个智能的Web导购系统。文中首次提出将交互式遗传算法和数量化I类理论这两种算法结合,进行系统在线学习并构建顾客的心理模型,使系统兼顾数量化I类理论的运算速度和遗传算法的非线性仿真模拟效果;同时为解决进化时间过长导致的用户疲劳问题,提出采用自主式GA来对用户历次所选的个体进行离线学习。通过对系统预测能力、运行速度的测试,以及用户对实验结果满意度的调查,证明这两种算法的结合弥补了原系统在非线性预测上的不足,使系统能够更快速、更准确、更人性化地追踪、模仿顾客的心理。Taken selling clothing for example,an intelligent shopping guide system is built in this paper,which is aimed to simulate user＇s mind and favor character.This paper proposes firstly combining Interactive Genetic Algorithm with Quantification Theory Ⅰ to implement system＇s customer psychological model construction and on-line learning.Because these two kinds of algorithms have their own strong points each, the new system possessed both Quantification Theory Ⅰ＇s speed and GA＇s non-linear simulate effect.At the same time,Independent Genetic Algorithms has been proposed to implement off-line learning using the pictures,which the user chosed before,to solve tiredness problem caused by the long time evolving process.Through examining the system＇s prediction ability,runtime,and investigating user＇s satisfaction degree to the experimental resuh,it has been proved that the combination of these two kinds of algorithms has remedied the deficiency of the original system in non-linear prediction,and has made the system track and imitate the customer＇s psychology faster,more exactly and humanized.

关 键 词：人工心理 智能导购系统 情感信息处理 遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]