管理工程学报2024，Vol.38Issue(3)：227-239,13.DOI:10.13587/j.cnki.jieem.2024.03.017
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
摘要
Abstract
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-Ⅲ algorithm.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-Ⅲ algorithm 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 directly decomposes a multilateral negotiation into multiple bilateral negotiations before solving the Pareto front,we first solve a public Pareto front for all the parties and then make the decomposition.This method maintains the nature of multilateral negotiation.On the one hand,solving the public Pareto front satisfies the situation in the actual multilateral negotiation,where resource allocation is tense and many mutually restrictive factors exist among multiple parties.On the other hand,multi-threaded negotiations satisfy the relative negotiation independence among different buyers.Finally,we verify the model's effectiveness in solving real-world problems through a case study. In the second part,we design two groups of experiments to test the model's validity.The first group examines the effect of improving the NSGA-Ⅲ algorithm on negotiation results.The second group studies the effect of the negotiation process on enhancing the efficiency of multilateral negotiation.First,the experimental results show that the improved NSGA-Ⅲ algorithm outperforms the NSGA-Ⅱ algorithm regarding the number of transaction rounds,the seller's utility,the buyer's utility,the utility product,and the utility difference.Therefore,the improved NSGA-Ⅲ algorithm can help buyers and sellers to obtain better utility and improve negotiation efficiency.Secondly,regarding the process design of multilateral negotiation,the experimental results of the transaction rounds number,the seller's utility,the utility product,and the utility difference in the multilateral negotiation process are significantly improved compared with the bilateral negotiation process.The experimental results show that the improved algorithm and negotiation process outperforms the latest research results in the field,thus effectively improving the efficiency of multilateral negotiation and helping achieve a win-win situation. In summary,we take one-to-many supply chain distribution as the research object and construct a one-to-many multilateral multi-issue negotiation model,which actively explores the problem of multilateral multi-issue automated negotiation in the supply chain environment.Through extensive experiments,we demonstrate the effectiveness of the improved NSGA-Ⅲ algorithm and the optimization of the multilateral negotiation process.From a practical point of view,the multilateral and multi-issue automated negotiation model designed in this paper is suitable for the transaction negotiation scenario between the upstream and downstream enterprises in the supply chain.It can effectively improve the negotiation efficiency between the seller and multiple buyers and helps to achieve a win-win situation for all parties.In addition,the example in this paper is limited to support one-to-many distribution scenarios.However,the proposed model and solution algorithm can be applied to broader one-to-many negotiation scenarios,such as one-to-many procurement.It can even be extended to scenarios of multiple buyers and multiple sellers,thus being the direction of future research efforts.关键词
供应链分销/多边多属性谈判/遗传算法/Pareto边界/AgentKey words
Supply chain distribution/Multilateral multi-issue negotiation/Genetic algorithm/Pareto front/Agent分类
信息技术与安全科学引用本文复制引用
曹慕昆,杨荇贻,党圣洁..面向供应链分销的多维空间Pareto边界自动谈判模型研究[J].管理工程学报,2024,38(3):227-239,13.基金项目
国家自然科学基金项目(72171199) The National Natural Science Foundation of China(72171199) （72171199）
