电子商务环境下考虑应急能力的大规模定制库存优化研究

E-commerce circumstances provide development opportunities for Mass Customization. Manufacturers display the optional models of product components on the e-commerce platform, from which customers could select each part to compose the custom products of their own preferences. Although Mass Customization under e-commerce circumstances could much better satisfied personalized demands, it might be difficult to be promoted in practice because of the complex inventory system and high inventory cost. The previous research on inventory optimization of Mass Customization assumed that the demands subjected to Poisson distribution which was with a known average rate and independently of the time since the last event. However, under e-commerce circumstances, because of the rapid spread of promotional information, influences of sales and evaluations, alternatives of components, and a lot of returns, the traditional assumption has no longer been valid. To solve above problems, we plan to use behavioral operations research and discrete system simulation as well as classical inventory theory and stochastic programming model, establish the expression of non-stationary fluctuant demands from behavioral research of customers, present inventory optimization and simulation models considering the emergency response capacity, and design the algorithms by intelligent optimization. Our research is expected to contribute to the optimization theory and method of complex multilevel stochastical inventory systems with non-stationary fluctuant demands, and provide scientific guidance for Mass Customization under e-commerce circumstances.

电子商务为大规模定制的发展提供了有利条件。制造商可以通过电子商务平台展示商品组成部件的可选型号，由客户选择组成个性化商品。尽管这种形式能够更好满足客户的个性化需求，但是复杂的库存系统和提高的库存成本使其在实践中推广面临巨大困难。大规模定制库存优化的已有研究建立在需求平稳性和无后效性假设上，然而电子商务环境下，由于促销信息快速传播、销量和评价的影响、零部件型号的可替代性、大量退换货等情况，使传统假设不再成立，需求的复杂相关性也增加了建模优化难度。针对上述问题，本项目拟在经典库存理论和随机规划模型基础上，综合行为运筹及离散系统仿真等方法，通过研究顾客行为，构建非平稳波动需求的数学表述，建立考虑应急能力的库存优化及仿真模型，结合智能优化设计求解算法。本项目旨在研究需求非平稳波动的复杂多级随机库存理论，探索解决复杂随机库存问题的新思路和新方法，为大规模定制借助电子商务平台的应用和推广提供科学指导。

电子商务环境下的顾客在线购物行为与传统店面相比有显著不同. 由于仅能通过网页介绍了解商品信息, 顾客在收到实物后, 经常发现与预期存在较大差异而选择退货. 此外促销信息的快速传播导致的顾客冲动购物、大量无缺陷退货可再次销售等情况也使得电子商务环境下的需求情况不再满足传统的平稳性和无后效性假设, 增加了大规模库存问题建模优化的难度. 针对上述问题, 本研究综合经典库存理论、行为运筹及系统仿真等方法, 通过研究顾客行为, 构建非平稳波动需求的数学表述, 建立库存和定价联合优化模型进行求解, 并进行仿真研究. 具体研究内容和主要结论包括：.从顾客购买前感知价值和收到实物后感知差异的角度, 构建了与价格相关的非平稳需求和退货率模型, 这一表述与已有实证研究中商品价格越高退货率越高的结论一致, 同时可较好解释大幅降价促销期过后高退货率的产生原因..对退货不再销售和可再销售的库存和定价问题进行联合优化建模, 进行求解, 分析最优解的特征.研究表明制定库存策略时考虑在销售期内再次出售退货, 可以用更低定价获取更大利润..考虑促销期顾客的冲动购买和退货审慎行为, 建立了日常销售期和促销期不同运费承担策略下库存和定价优化模型, 对不同情形下的运费承担策略进行比较..采用离散系统仿真和系统动力学方法建立了非平稳需求下的大规模定制库存仿真模型, 分析得出尽管定制产品提高了库存总成本和库存水平, 但是可以获得更高的服务水平..本项目研究丰富了需求非平稳波动的随机库存理论, 探索了解决复杂随机库存问题的新思路和新方法, 为电子商务环境下的定价库存等问题提供了决策支持.