基于遗传算法和BP算法的比较购物Agent模型.pdfVIP

ISSN 1000翩 1239!CN 11翩 1777!TP 计算机研究与发展 Joumal of Computer R部earch and Development 43(Supp1.); 51-55 , 2006 基于遗传算法和 BP 算法的比较购物 Agent 模型 郑盛福l 胡山立1 ， 2 林超峰1 苏射雄1 (福州大学计算机科学与技术系 福州 350002) (中国科学院计算机科学撒点实验裳 北京 100080) (sfz_ roger@ yah∞·∞) AnAgentM创.el B凶ed on Genetic Algorithm and BP Algorithm in Comparison Shopping Zheng Shengfu1 • Hu Shanli1,2 , Lin Chaofen矿， and Su Shexiong1 l( IJeμrtment of Comρuter Science and Technology , Fuzhou University ， 归纳ou 350002) 2 (Laboratory for Com如ter Scien饵， Chinese Academy of Sciences. Beijing 100080) Abstract First-generation shopping agents provide great convenience for consumers. However , they are limited to comparing merchant offerings only on price. To cope with this , this paper proposes an agent in comparison shopping. which is based on a mixed neural network algorithm composed by genetic algorithm and BP algorithm. It can search and filtrate the required informatíon according to the preference of customers and the searching ex阳rience of others. After the search , it arranges the suitable g∞ds based on the preference of customers , so that customers can buy the well-content goods without browsing plenty of web sites. Simulation índicates that this model is feasible and effective , being able to predict the preference of customers , etc. Key words genetic algorithm; BP algorithm; comparison shoppingj electronic ωmmercej agent 摘要第一代购物Agent 给消费者带来极大的便利，但是它们只是局限于商品价格的比较.针对这个 问题，提出一种暴于遗传算法和日P 算法的比较购物 Agent 模型，它可以根据消费者的偏好和别人的搜 索经验为消费者搜索和过滤所常的信息，并且能在搜索先后，将符合妥求的商品按消费者的偏好进行排 序，使消费者不常要浏览大寰的网站就能买到称心如意的商品.仿真实验表明，该模型是行之有效的， 它能准确地预圳消费者的偏好等. 关键词遗传算法jBP 算法;比较购物;电子商务j Agent 中回法分类号 TP18 Agent 技术的电子商务.近年来，许多学者和专家试 随着电子技术和 Internet 的发展，信息技术作