面向供应链分销的多维空间Pareto边界自动谈判模型研究  

An automated negotiation model based on multi-dimensional space pareto front for supply chain distribution

作　　者：曹慕昆[1] 杨荇贻 党圣洁 CAO Mukun;YANG Xingyi;DANG Shengjie(School of Management,Xiamen University,Xiamen 361005,China)

机构地区：[1]厦门大学管理学院,福建厦门361005

出　　处：《管理工程学报》2024年第3期227-239,共13页Journal of Industrial Engineering and Engineering Management

基　　金：国家自然科学基金项目(72171199)。

摘　　要：随着电子商务的快速发展,自动谈判逐渐成为提升供应链系统效率的一种手段。为了优化多方参与的供应链分销谈判应用,本文将多边多属性谈判问题转化为多目标优化模型,采用改进的非支配遗传算法NSGA-Ⅲ计算多维空间的Pareto边界;然后,设计多线程谈判模型,将参与多方谈判的买卖各方拆解为多个双边谈判线程,分别在多维Pareto边界上进行谈判;进而,采用动态时间依赖策略(DTD),使Agent根据对方报价在Pareto边界上动态调整让步策略,快速达成协议。为验证模型的有效性,本文进行了大量模拟自动谈判实验。实验结果表明,所提出的改进算法和谈判流程优于领域最新研究成果,能有效提升多边多属性谈判效率,有助于多方达成共赢局面。With the rapid development of e-commerce,the negotiation between enterprises has gradually shifted from offline to online,which has created new demands and challenges for intelligent negotiation technology.For example,along with the digitalization of the supply chain system,supply chain management has undergone some new changes,such as supply chain distribution.Therefore,it is imperative to seek intelligent negotiation techniques to optimize supply chain distribution decisions and improve the efficiency of supply chain distribution systems.To optimize the application of automated negotiation in supply chain distribution,we transform the multilateral multi-issue distribution negotiation problem into a multi-objective optimization model and conduct the multilateral multi-issue negotiation on the Pareto front in a multi-dimensional space.In the first part,we introduce the construction of the automated negotiation model for the supply chain distribution problem from five aspects in detail.They are the negotiation element model establishment,the solution algorithm design,the negotiation process design,the negotiation mechanism design,and the practical case application.Firstly,we build a multi-objective optimization model for the supply chain distribution problem and solve the multi-objective optimization model based on the improved NSGA-Ⅲalgorithm.By eliminating the equivalent solutions in the population evolution process,the improved non-dominated genetic algorithm NSGA-Ⅲalgorithm can better maintain population diversity when dealing with high-dimensional problems.Secondly,after using the improved NSGA-Ⅲalgorithm to obtain the multi-dimensional public Pareto front,we decompose the one-to-many multilateral negotiation into multiple one-to-one bilateral negotiations.Thirdly,following the designed negotiation process and mechanism,the multi-party agents adopt the dynamic time-dependent strategy(DTD)to start multi-threaded bargaining on the multi-dimensional public Pareto front.Unlike the previous work that directl

关 键 词：供应链分销 多边多属性谈判 遗传算法 Pareto边界 AGENT 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]