一种基于信任的多Agent供应链协商关系网

Agent技术已被广泛用于供应链伙伴的协商。协商前如何选择协商Agent对提高协商效率有着重要的意义。提出了一种基于信任的多Agent协商关系网及其形成和更新算法,并对该协商关系网的特点进行了深入的研究。模拟表明,提出的协商关系网能有效地促进Agent之间的协商,提高协商成功率。     

一种劝说式多Agent多议题协商方法

多Agent系统中的协商问题往往由许多议题组成，导致问题空间十分庞大．传统的协商方法通过对问题空间进行穷尽搜索来找到最优解，并不适合多议题协商．而且，传统的方法不考虑协商偏好变化的情况，使得Agent在不完全及不正确环境下找到的最优解并不合理．提出一种劝说式多Agent多议题协商方法．借助信念修正这一有效的推理工具，协商Agent能够在协商过程中接受协商对手的劝说，考虑对手对协商议题的偏好，并根据一种基于辩论的信念修正方法调整自身的偏好．这样就能够使协商Agent对变化的协商环境具备适应性，从而提高协商的效率及正确率，快速准确地达成协议．     

Agent劝说中的积极舆论模型及其评价

Agent劝说中的舆论能发挥其人工智能优势,同时节约谈判成本和提高谈判效率。为此,给出Agent劝说中舆论的定义,并将Agent劝说中的舆论分为积极舆论和消极舆论,在定义和分类的基础上,结合形式逻辑理论,提出Agent劝说中的积极舆论模型,并通过相应的举例对模型进行解释。针对该模型,提出相应的评价体系,并通过相应的算例对该体系进行验证。结果表明,该模型和评价体系能促使Agent劝说进一步快速高效完成。     

基于拍卖博弈的电子商务自动协商

为提高电子商务自动协商系统效率,本文以拍卖博弈理论为基础,提出并实现了一种基于拍卖博弈的自动协商Agent模型,并在此基础上实现了一个基于拍卖博弈的电子商务自动协商原型系统,应用于一个企业敏捷供应链管理系统中实现自动协商交易。     

基于博弈分析的电子商务自动协商系统

为提高基于拍卖机制的电子商务自动协商系统效率，该文以非合作博弈论为基础提出了基于博弈分析的自动协商Agent模型及基于拍卖机制和博弈分析的自动投标协商算法，实现了一个基于拍卖机制和博弈分析的电子商务自动协商原型系统，并应用在一个企业敏捷供应链管理系统中实现自动协商交易。     

基于Agent正面情绪变化的劝说及评价模型

Agent劝说在提高电子商务谈判自动化程度的同时也能相应提高其理性程度,因此受到广泛关注。Agent情绪变化则能影响劝说的进程和效果。对Agent劝说中的情绪变化进行分类;结合形式化理论,提出了基于Agent正面情绪变化的劝说模型,并对模型进行举例说明;为使该劝说模型有效,提出了Agent正面情绪变化度的概念;在此基础上,构建了相应的评价模型;最后通过算例对模型的有效性进行了验证。     

随机需求下供应链企业订购多Agent协商模型

针对随机需求下供应链产销订购冲突问题,考虑一个制造商和两个零售商组成的两级供应链,构建主从博弈下的多Agent协商模型。制造商作为博弈主方制定批发价,零售商作为从方选择最优订货量和零售价,制造商Agent和零售商Agent自动协商,运用模拟退火算法寻求模型的最优解。通过算例发现合理的让步策略和收益共享契约能够提高供应链系统利润,实现产销双方的互利共赢。验证模拟退火算法求解该模型比遗传算法能够得到更优解。     

基于冲突检测的供应链协同计划

在供应链协同计划中,以供应链整体最优为目标做出的决策有可能引发个体理性和集体理性冲突。针对该问题,在研究供应链协同计划时引入冲突检测方法,建立供应链协同计划冲突检测模型,采用Agent协商技术对检测到的冲突进行消解。仿真实例说明,在供应链协同计划中引入冲突检测方法能及时发现供应链中存在的冲突,有助于提高供应链协同计划的效率和科学性。     

基于案例匹配度的Agent劝说机制

自动谈判中的Agent劝说能充分提高其在动态环境下的谈判效率,保证谈判效果,同时极大节约谈判成本。案例推理技术能使Agent将谈判中遇到的新案例与以往成功案例进行比对,根据结果做出较优的选择,从而促使Agent之间取得更佳的合作效果。对Agent劝说中的案例匹配度的概念进行了界定;结合形式逻辑理论,建立了基于案例匹配度的Agent劝说模型,并通过举例对模型进行了阐述;提出了模型中关键指标的计算方法,并通过算例进一步给出了相应的评价体系和交互机制;最后进行了验证,实现了基于案例匹配度的Agent劝说机制的构建。实验结果表明,该机制能有效减少系统开销。     

基于学习的多Agent多议题协商优化研究

以买方Agent的观点,对从交易平台上获得的卖方Agent的历史协商信息进行分析,并根据其特点做初步过滤。在此基础上,针对现有协商模型中存在的问题,提出了一个Agent协商对手选择算法和相应的交互机制,并验证了其可行性。该算法可用于Agent协商开始前协商对手的选择和初始信念的更新,对Agent在协商中策略的选择和执行具有指导作用,能有效提高Agent在协商中的效用及效率。     

制造业供应链产销协同计划冲突协商模型

面对市场竞争日益剧烈和客户需求多样化的趋势,制造业供应链的制造商和经销商努力实现产销协同计划。然而在产销协同计划中时常出现冲突,及时有效消解冲突,能提高整个供应链的协作效率,改善供应链上企业间的合作关系;反之,会降低供应链的运作效率,削弱供应链的竞争力。针对这类冲突问题,引入让步协商策略,在有限信息共享条件下,建立供应链产销协同计划冲突协商模型;设计具有历史提议回顾特点的协商流程;通过文化基因算法,产生反提议生成策略;通过算例验证文化基因算法及冲突协商模型的有效性。     

Formulation and analysis of dynamic supply chain of backfill in construction waste management using agent-based modeling

Backfill is the excavated material from earthworks, which constitutes over 50% of the construction wastes in Hong Kong. This paper considers a supply chain that consists of construction sites, landfills and commercial sources in which operators seek cooperation to maximize backfill reuse and improve waste recovery efficiency. Unlike the ordinary material supply chain in manufacturing industries, the supply chain for backfill involves many dynamic processes, which increases the complexity of analyzing and solving the logistic issue. Therefore, this study attempts to identify an appropriate methodology to analyze the dynamic supply chain, for facilitating the backfill reuse. A centralized optimization model and a distributed agent-based model are proposed and implemented in comparing their performances. The centralized optimization model can obtain a global optimum but requires sharing of complete information from all supply chain entities, resulting in barriers for implementation. In addition, whenever the backfill supply chain changes, the centralized optimization model needs to reconfigure the network structure and recompute the optimum. The distributed agent-based model focuses on task distribution and cooperation between business entities in the backfill supply chain. In the agent-based model, decision making and communication between construction sites, landfills, and commercial sources are emulated by a number of autonomous agents. They perform together through a negotiation algorithm for optimizing the supply chain configuration that reduces the backfill shipment cost. A comparative study indicates that the agent-based model is more capable of studying the dynamic backfill supply chain due to its decentralization of optimization and fast reaction to unexpected disturbances.     

基于多Agent供应链协同的自适应协商方法

在供应链协同过程中常出现价格、数量、交货时间等冲突,若不及时消解会影响供应链的整体利益。为有效化解供应链协同冲突,弥补传统协商缺点,提出一种基于多Agent的自适应协商方法。该方法以二级供应链,即制造商和供应商的订货为研究对象,使用多Agent供应链协同作为约束条件,采用案例推理为主要协商算法,引入灰色关联度到案例集与目标案例相似度中,将遗传算法应用于相似案例议题权重的优化中。通过算例验证该方法可简化案例相似度的计算,提高解决冲突的效率,并加强自适应性,为消解冲突提供最优决策。     

Development and evaluation of the many-to-many supplier negotiation strategy

Supply chain is a complex logistics network, consists of suppliers, manufacturers, warehouses, distribution centers, retailers and customers. In general, the number of firms in each supply chain tier is more than one. Thus, the supply–demand relationship of two adjacent tiers in supply chain is the many-to-many type. Most of supplier selection processes are based on bidding and negotiation mechanism. Due to the diversity of intellectual backgrounds of the negotiating parties, competitive nature, many variables involved in supply–demand relationship, complex interactions and inadequate negotiation knowledge of project participants, the supplier negotiation process is sophisticated. This paper intends to develop the many-to-many supplier negotiation process, design the interactive and competitive bidding strategies of both-side parties, and conduct experimental simulations to investigate how the negotiation efficiency influenced by the varying bidding strategies of a considered demander. The purpose of this paper is to explore the relationships between a demander’s bidding strategies and negotiation efficiency under different order and competitive conditions. The explorative results will be the decision support for assisting a demander to determine the proper bidding strategy in many-to-many supplier negotiation process.     

基于软件人的装备物流供应链协商模型

供应链成员之间的协商是智能供应链管理系统的关键部分,供应链成员只有通过协商才能进行有效的交互、合作。装备物流供应链中的一些运作功能定义成多个软件人模型,其中包括协调软件人、采购软件人、生产软件人、库存软件人、运输软件人、供应软件人。本文主要研究的是采购软件人与供应软件人之间的协商,提出一个软件人之间的协商模型。     

Agent-based web service for the design of a dynamic coordination mechanism in supply networks

The theory of network coordination presents an effective approach to improve the business processes within supply networks. The automation of the negotiation process among buyers and suppliers has become an important policy in the transactional networks. This leads to assessing the roles of both quantifiable and non-quantifiable parameters in coordination mechanisms with the aim of achieving higher performance. Here, we develop an e-based supply chain multi-agent model for the design of mass-customized on-line services. The model addresses the bullwhip effect in multi-stage supply chain and also clarifies the evaluation of inventory policies in various supply and demand uncertainties. To illustrate the feasibility of the approach, we implement a prototype system and evaluate its performance by simulation using Colored Petri Nets (CPNs). The validation results reveal the model efficiency in providing a more realistic optimization process that takes the dynamic information flow in uncertainty environments into consideration.     

The impact of product characteristics and innovativeness on the benefits of collaboration

Horizontal collaboration is a promising avenue to improve the efficiency of logistical operations. However, the benefits strongly depend on the degree of fit between partners. In this paper, we analyze the impact of the partners' product characteristics on those benefits, focusing on their innovativeness. Companies supplying functional versus innovative products have different requirements in supply chain efficiency and responsiveness, which impacts the benefits that can be reached with a given partner. To assess the collaborative benefits, we use a location–inventory model accounting for the partners' individual interests and the costs revealing the responsiveness level of the supply chain (facilities, transportation, cycle inventory, safety stocks and stock-outs). The model offers a set of Pareto-optimal solutions balancing the partners' costs to support the selection and negotiation process. Finally, we perform numerical experiments in which the partners supply products with identical or different levels of innovativeness and with various demand volumes, leading to valuable managerial insights on the impact of product characteristics on collaborative benefits.