电子商务自动协商研究-计算机应用技术专业论文.docxVIP

重庆大学工程硕士学位论文中文摘要 重庆大学工程硕士学位论文 中文摘要 I I 摘 要 目前，电子商务自动协商技术已经成为电子商务有效实现中涉及的一个重要 问题。但现在尚无一种令人满意的、具有学习能力的、对现实的商务协商进行有 效的形式化的模型。 针对这一问题，本文提出基于神经网络的协商经验学习与推理的自动协商思 想，即利用神经网络与数据库技术对协商经验进行学习与管理，有效的执行协商 策略，更好的对现实商务交易行为进行模拟，实现分布式、动态电子商务环境中 的智能协商。 在此基础上，本文基于 Agent 建立了电子商务自动协商的逻辑模型，并采用 Java 语言设计实现了模型的基本功能。在该模型中，建立了基于神经网络的协商 决策函数、协商经验学习与推理、经验库的建立，实现根据协商环境变化动态选 择协商策略的商务智能；用 Agent 技术建立了多属性自动协商模型，并对模型进 行试运行，研究实际自动协商效果。 基于神经网络可学习自动协商研究结果表明：（1）用神经网络构建协商决策 函数可以灵活、全面、真实的代表用户的利益；（2）基于神经网络的协商经验学 习与推理可以利用人类以往的协商经验，促使新交易的成功；（3）用 Agent 技术 建立的多属性自动协商模型更加接近实际，实现了分布式、动态电子商务环境中 的智能协商。 关键词：自动协商，神经网络，Agent，电子商务 重庆大学工程硕士学位论文英文摘要 重庆大学工程硕士学位论文 英文摘要 II II ABSTRACT Nowadays, auto-Negotiation in E-commerce has become an important issue while taking the E-commerce into reality , but there is still no satisfying and learning ability solution model to formalize the process of commerce negotiation . On that account , this paper presented an auto-Negotiation mechanism that can learn and deduce experiences based on neural networks . To get the successful emulation of the real commerce behavior with effective negotiation strategy , it is suggested that we must take full advantages of experiences learned and organized by neural networks and data bases while realizing the intelligent negotiation in distributed and dynamic E-commerce environment . Therefore, we construct a logic model for E-commerce auto-Negotiation and also put its basic functions into an executable program with Java programming language. In this model , we construct the decision-making function of negotiation based on neural networks ,and method of experiences learning and deducing ,and experience bases ,and realize commerce intelligence with negotiation strategy on environment . We also construct more attributive auto-Negotiation model and run the model to test result . The results of learnable auto-negotiation by neural networks show that : (1) It is flexible perfect and real to represent user’s benefit by decision-making function of negotiat