主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2024, Vol. 32 ›› Issue (2): 54-64.doi: 10.16381/j.cnki.issn1003-207x.2021.0339

• • 上一篇    下一篇

不完备质量下带有风险厌恶的库存决策模型

陈杰1,邢灵博1(),李胃胜1,陈志祥2,陈崇萍3   

  1. 1.海南热带海洋学院理学院, 海南 三亚 572022
    2.中山大学管理学院, 广东 广州 510275
    3.华南师范大学政治与公共管理学院, 广东 广州 510006
  • 收稿日期:2021-02-22 修回日期:2022-04-22 出版日期:2024-02-25 发布日期:2024-03-06
  • 通讯作者: 邢灵博 E-mail:qzuxinglingbo@163.com
  • 基金资助:
    国家自然科学基金项目(72061011);海南省自然科学基金高层次人才资助项目(720RC662);广东省自然科学基金项目(2020A1515111044);广东省哲学社会科学“十三五”规划项目(GD20YGL05)

A risk-Averse Inventory Decision-making Model with Quality Incompleteness

Jie Chen1,Lingbo Xing1(),Weisheng Li1,Zhixiang Chen2,Chongping Chen3   

  1. 1.School of Science, Hainan Tropical Ocean University, Sanya 572022, China
    2.School of Business, Sun Yat-sen University, Guangzhou 510275, China
    3.School of Politics and Public Administration, South China Normal University, Guangzhou 510006, China
  • Received:2021-02-22 Revised:2022-04-22 Online:2024-02-25 Published:2024-03-06
  • Contact: Lingbo Xing E-mail:qzuxinglingbo@163.com

摘要:

在质量的不完备性诱发联动效应的环境下,考虑带有风险厌恶的库存决策问题。假设质量水平为一马尔可夫过程,并将马氏理论和条件风险值(conditional value-at-risk, CVaR)准则纳入库存决策模型的框架,进而给出了不完备质量下带有风险偏好的决策模型及其最优决策机制。模型数值实验的结果表明:最优期望订购量和期望销售利润关于供应能力和风险厌恶因子都为单调递增;库存系统的可靠性与随机质量过程的不可约性、遍历性等基本性质有关,即当随机质量过程为不可约的遍历马氏链时,模型的最优解具有平稳的极限性态。

关键词: 不完备质量, 马尔可夫链, CVaR准则, 库存决策

Abstract:

Quality uncertainty is one of the most important risk sources of the inventory system, which has a negative correlation with the supply capacity on the supply side. Meanwhile, the corresponding linkage effect induced by the quality uncertainty has a disturbing effect on the decision-making behavior of the inventory system, and then affects the risk preference of decision-makers. Therefore, based on the uncertainty of supply capacity, purchase price, wholesale price, residual value and other factors caused by quality imperfection, an inventory decision model with risk aversion under imperfect quality is proposed by Markov process. According to the results of the new model and numerical simulation, the following important conclusions and management implications can be drawn:Firstly, the effect of quality level and supply capacity on optimal expected order quantity and expected profit of inventory system has consistent attribute, that is, the effect of quality level and supply capacity on both have a positive correlation. When the quality level and supply capacity are in the optimal state, the optimal expected order quantity and expected profit reach the optimal value, otherwise vice versa. In addition, if the quality process{Zk} meets ergodicity and irreducibility, the inventory system has a good robustness, and the decision maker can obtain relatively stable expected profits as long as he takes a robust ordering strategy.Secondly, under the multiple random decision environment which is formed by the quality level, supply ability, pleased, wholesale price, the salvage value and demand, the transmission mechanism of risk aversion factor has an invariance property to the decision making mechanism, that is, the attribute of relationships (positive correlation) between risk aversion factor η and optimal expected order quantity and expected profit is not affected by exogenous random factors. This further verifies an important conclusion in behavioral science, that is, in different decision-making environments, the higher the degree of risk aversion of decision makers, the more conservative they will adopt strategies.Thirdly, the relevant conclusion shows that the post-decision information set has certain advantages which is derived from incorporating Markov process into the theoretical framework of the classic risk-averse newsboy model, that is, based on the statistical regularity of Markov process, the traditional post-decision information can be extended from two dimensions to five, and then enrich the ideological connotation of the post-decision information set, to describe the movement mechanism for the optimal expected order quantity, expected profit, order quantity robustness, expected profit stability, return expected profit and other elements of the inventory system. Obviously, Multi- dimensionalization of post-decision information set is more conducive to revealing the statistical structure of stochastic inventory system and perfecting the evaluation mechanism of reliability and performance of inventory system.

Key words: imperfect quality, Markov chain, CVaR criterion, inventory decision

中图分类号: