基于Agent的多问题并行遗传谈判模型

Agent谈判增加了电子商务系统的主动性,一个有效的谈判模型是系统实现的关键。分析了已有Agent谈判模型的特点,设计的基于Agent的多问题并行谈判模型解决了已有谈判模型中存在的谈判问题单一、非并行、不考虑对手收益和固定权重等问题。模型中产生谈判方案的算法的自适应性体现在问题权重、遗传参数和收益偏差的动态调整上,给出了问题实数编码和权重调整公式。最后,设计了一个面向三个问题的电子谈判实例,验证了谈判模型的可行性和有效性。

Multi-Issues and Parallel Genetic Negotiating Model Based on Agent

Agent negotiation improves e-business system activity,A valid negotiation model is the key to realize the system.This paper designs a multi-issues and parallel genetic negotiation model based on agent after analyzing the features of agent negotiation existed in current models,It solves the problems of single issue,not parallel,no consideration to opponent gains and fixed weight in these models.The self-adaptive property of algorithm producing negotiation schemes in model embodies in adjustment dynamically on issue weight,genetic parameter and bias of gains, and gives the formulas of issue weight adjustment and real encoding.Lastly,an e-negotiation instance faced three issues is designed,it manifests its practicability and validity.