基于风险厌恶的多市场报童鲁棒优化模型与策略研究

发布时间：2019-04-09 13:27

【摘要】：由于在现实经济生活中的普遍性,报童模型一直是供应链管理,特别是易逝品供应链库存管理领域的热点。面对产品较快的更新换代速度和消费者的个性化需求,短生命周期产品的市场需求面临着很大的不确定。传统研究在处理不确定性时通常假设其服从某一已知分布,通过优化期望利润或成本方式获得最佳决策。然而,实证研究表明,决策者在实际运作中采取的决策并不总是与基于期望利润或成本的方法得到的最优策略一致。在面对不确定性需求时,决策者往往呈现一定的风险态度,从而影响最终决策。此外,随着信息技术的飞速发展以及交通运输等物流系统的日益完善,企业的服务对象不再局限于单一市场,对多个市场的集成考虑为企业获取更多的利润提供了机会。基于此,本文在需求信息不确定条件下,研究了具有风险厌恶的多市场报童鲁棒优化问题。研究结果将为面向多个市场销售产品或服务的企业决策者应对需求不确定性扰动提供决策支持。本文的主要研究内容如下：首先,针对离散需求情景概率分别隶属于区间和椭球不确定集情况,建立了基于期望利润最大化的鲁棒报童模型,给出了需求情景概率不确定条件下的鲁棒对应。分别运用线性规划和拉格朗日对偶理论,将区间和椭球不确定集下的鲁棒对应模型转化为易于求解的线性规划和二阶锥规划问题。进一步,考虑决策者对不确定性导致的绩效风险厌恶态度,以条件风险值(CVaR)刻画的绩效作为决策目标函数,研究两种不确定集下的鲁棒报童模型。其次,在上述研究基础上,考虑决策者同时面临多个市场,分别建立了基于期望利润和条件风险值的鲁棒多市场报童模型。针对不确定需求情景概率分别隶属于区间和椭球不确定集情况,运用对偶理论将问题转化为易于求解的线性规划和二阶锥规划问题。最后,对所建模型进行了数值计算与分析,验证了通过文中方法求出的鲁棒策略的有效性。结果表明,同需求真实分布的最优情况相比,本文提出的鲁棒决策虽然会导致系统绩效损失,但损失值很小,说明根据文中所给方法求解的鲁棒策略具有良好的鲁棒性,能够有效抑制需求分布不确定性对系统绩效的影响,绩效损失值可以解释为决策者为了获得真实需求分布信息所愿意支付的最高成本。此外,进一步分析了决策者风险厌恶程度变化对鲁棒决策的影响,结果表明随决策者风险厌恶程度的增加,订货量和系统绩效呈减少趋势。
[Abstract]:Due to its universality in real economic life, newsboy model has always been a hot spot in supply chain management, especially in the area of inventory management of perishable goods supply chain. In the face of the rapid updating speed of the product and the individual demand of the consumers, the market demand of the short life cycle product is faced with a great deal of uncertainty. When dealing with uncertainty, the traditional research usually assumes that it obeys a certain known distribution and obtains the best decision by optimizing the expected profit or cost. However, empirical research shows that the decision-making taken by decision-makers in practical operation is not always consistent with the optimal strategy based on expected profit or cost-based approach. In the face of uncertain demand, decision makers often present a certain attitude of risk, thus affecting the final decision-making. In addition, with the rapid development of information technology and the increasingly perfect logistics system such as transportation, the service object of an enterprise is no longer limited to a single market, and the integration of multiple markets provides an opportunity for enterprises to obtain more profits. Based on this, this paper studies the multi-market newsboy robust optimization problem with risk aversion under the condition of uncertain demand information. The results will provide decision support for decision makers who sell products or services in multiple markets to deal with the uncertainty of demand. The main contents of this paper are as follows: firstly, a robust newsboy model based on expected profit maximization is established for the probability of discrete demand scenarios belonging to interval and ellipsoidal uncertain sets, respectively. The robust correspondence under uncertain probability of demand scenario is given. Using linear programming and Lagrangian duality theory, the robust corresponding models under interval and ellipsoidal uncertain sets are transformed into linear programming and second-order cone programming problems which are easy to solve. Furthermore, taking the performance described by conditional risk value (CVaR) as the decision-making objective function, the robust newsboy model under two uncertain sets is studied, considering the attitude of decision makers to performance risk aversion caused by uncertainty. Secondly, a robust multi-market newsboy model based on expected profit and conditional risk is established based on the above-mentioned research, considering that the decision maker faces multiple markets at the same time. Considering that the probability of uncertain demand scenario belongs to interval and ellipsoidal uncertain set respectively, the dual theory is used to transform the problem into linear programming and second-order cone programming, which are easy to solve. Finally, numerical calculation and analysis of the model are carried out to verify the effectiveness of the robust strategy obtained by the method in this paper. The results show that, compared with the optimal demand distribution, the proposed robust decision can result in loss of system performance, but the loss value is very small, which shows that the robust strategy based on the method presented in this paper has good robustness. It can effectively restrain the influence of uncertainty of demand distribution on system performance, and the value of performance loss can be interpreted as the maximum cost that the decision maker is willing to pay in order to obtain the real demand distribution information. In addition, the influence of the change of risk aversion degree on the robust decision-making is further analyzed. The results show that with the increase of the risk aversion degree of the decision-maker, the order quantity and system performance tend to decrease.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F274

【相似文献】