中文摘要：

为探讨顾客需求变动下冷链品配送路径的规划与优化模型选择问题，首先以带时间窗的车辆路径问题模型为基础，构建了以配送总成本最小化为目标的冷链品配送模型；然后，通过节约算法进行配送路径的优化；最后，通过算例比较了运用随机性配送模型（考量顾客需求的变动）与确定性配送模型（使用给定需求期望值）所规划的优化路径及其差异。结果显示，当顾客需求波动幅度较小时，确定性模型规划优于随机性模型，当顾客需求波动幅度较大时，随机性模型规划优于确定性模型；此外，随着需求变动程度的增大，两种模型所得出的配送总成本都会相应增加，因此，最后提出了降低顾客需求不确定性的相关建议。

英文摘要：

To discuss the planning and selection of optimization model for distribution routing of cold chain items with variable demand of customer, first, we established a distribution model of cold chain items with the objective of minimizing the total distribution cost based on the model of vehicle routing problem （VRP） with time window. Then, the distribution routing is optimized by saving algorithm. At last, we compared the optimal routes planed by the stochastic distribution model （considering demand variation of customer） and the deterministic distribution model （based on a given demand expected value） and their differences by a numerical example. The result reveals that （1） the planning by the stochastic model is superior to the deterministic model when demand fluctuates wildly and vice versa; （2） the total distribution costs obtained by the 2 models will increase with the increase of the degree of changes in demand. Therefore, we proposed some suggestions to reduce the demand uncertainty of customer.