供应链配货博弈决策行为研究

发布时间：2018-07-05 01:37

本文选题：订货决策 + 配货博弈　；参考：《清华大学》2014年博士论文

【摘要】：供应短缺中的决策本质上可归结为一个博弈问题，经典研究基于纳什理论已有一些成果。然而既有的研究范式存在着一个基本性的局限，体现在决策模型中对人的行为模式做了极简化假设，认为决策者是“完全理性”人，具有完美的信息获取和处理能力，这与实践中人的管理决策存在着差异，进而限制了其研究成果在管理实践中的应用。针对上述问题，本文提出在配货博弈中引入对行为决策的考量，以零售商的订货决策为切入点，通过理论和实验相结合的方法，，开展行为决策研究。论文系统地考察了需求信息完全和不完全两种博弈类型、按比例分配和小订单优先两种规则构成的四类配货博弈问题。对完全信息配货博弈的研究表明，被试者订货决策与纳什均衡解存在系统性偏差，且被试者在重复决策中具有纳什理论所未能刻画的学习效应。本文提炼了不同分配规则、不同运作成本对被试者的订货决策和供应链运作绩效的影响规律。进一步，通过定量分析，提出并建立了基于随机最优均衡的行为决策模型，揭示了被试者在博弈过程中的有限理性和学习效应，模型预测结果实现了对实验数据的准确逼近。对不完全信息配货博弈的研究，发现被试者不仅存在博弈的有限理性，还对不同运作成本持有差别化的估计。本文基于心理账户理论和随机最优均衡，提出并建立了相应的行为决策模型；通过定量分析，揭示了被试者在不同分配规则、不同成本和不同轮次中对缺货成本和浪费成本持有的心理账户特点，以及系统运作效率的特点，并实现了预测结果对实验数据的准确逼近。通过在两类博弈中开展的分配规则比较研究，指出了现有研究尚为薄弱的小订单优先的优点和特色：与按比例分配相比，能够诱导真实人制定更贴近标准理论的订货决策，供应链系统对运作成本的波动具有更强的稳定性，并且能够获得更高的运作绩效，是很多应用场合中值得选用的一种分配规则。本文建立了在配货博弈中考察行为决策的研究框架，提出了有效的行为决策模型，为今后进一步开展供应链优化管理和机制设计研究奠定了基础。
[Abstract]:Decision making in supply shortage is essentially a game problem. Some achievements have been made in classical research based on Nash theory. However, the existing research paradigm has a basic limitation, which is reflected in the extremely simplified assumption of human behavior model in the decision-making model, which holds that the decision-maker is a "completely rational" person with the perfect ability to obtain and process information. This is different from people's management decision in practice, which limits the application of their research results in management practice. In order to solve the above problems, this paper proposes to introduce the consideration of behavior decision in the distribution game, take the ordering decision of the retailer as the starting point, through the combination of theory and experiment, carry out the behavior decision research. In this paper, four kinds of distribution game problems, which are composed of two game types of complete and incomplete demand information, proportional distribution and small order priority, are systematically investigated. The study on the game of complete information distribution shows that there is a systematic deviation between the ordering decision and Nash equilibrium solution, and that the subject has the learning effect that Nash theory can not depict in repeated decision making. This paper abstracts the influence of different allocation rules and different operating costs on the order decision and supply chain performance of the subjects. Furthermore, through quantitative analysis, a behavioral decision-making model based on stochastic optimal equilibrium is proposed and established, which reveals the limited rationality and learning effect of the subjects in the game process. The prediction results of the model achieve the accurate approximation of the experimental data. Based on the study of incomplete information distribution game, it is found that the subjects not only have the bounded rationality of the game, but also have different estimates of different operating costs. Based on the theory of psychological account and stochastic optimal equilibrium, this paper proposes and establishes the corresponding behavioral decision model, and through quantitative analysis, reveals the subjects in different allocation rules. The characteristics of psychological accounts and the operating efficiency of the system in different cost and different batches are obtained and the predicted results are accurately approximated to the experimental data. Through the comparative study of distribution rules in two kinds of games, this paper points out the advantages and characteristics that the existing research is still weak in small order priority: compared with proportional distribution, it can induce real people to make order decision which is closer to the standard theory. The supply chain system has stronger stability to the fluctuation of operating cost and can obtain higher operation performance. It is a kind of allocation rule that is worth choosing in many applications. In this paper, the research framework of behavior decision in distribution game is established, and an effective behavioral decision model is put forward, which lays a foundation for further research on supply chain optimization management and mechanism design.
【学位授予单位】：清华大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：F274;F224.32

【参考文献】