基于“时间-温度”流数据分析的生鲜冷链配载计划逆优化方法研究

The characteristics of cold chain logistics of perishables include small-batch, multi-batch, multi-temperature co-distribution and different temperature sensitivity of fresh food, which cause the irreconcilable difficulties in balancing the distribution cost and quality assurance. Fresh food logistic companies often optimize the distribution plan with the goal of minimizing costs. However, practice has proved that it is difficult to sustain low costs at the expense of food quality in the long run. The ability to improve the quality of fresh food transportation within a given cost threshold has become a key problem that affects people's livelihood and enhances the core competitiveness of logistics companies. We found that "time-temperature" stream data is the key to solve this problem. Therefore, the project proposes an "inverse optimization" academic idea, which refers to maximize the freshness through temperature adjustment within a given logistics cost threshold. The project focuses on the introduction of quantitative prediction of shelf-life based on the "time-temperature" stream data into distribution planning model. The project achieves an integrated solution of ""time-temperature" stream data mining → cost-minimization original distribution model → dual conversion of inverse optimization → intelligent-based optimization algorithm ". Besides, we conduct application research with a large fresh food logistics company in Shenyang. Instead of solely cost control theory, this project adopts a food-safety perspective of temperature control to optimize the fresh food distribution plan, which change from"cost-driven" to"quality-driven" . The proposal provides a new theory and method for the quality improvement of fresh food logistics. Also, it is a theory and application extension of inverse optimization in the field of fresh logistics.

生鲜冷链的小批量、多批次、多温共配及各生鲜品类最佳仓储温度的相异性，导致配载成本与食品品质难以兼顾。企业常以成本最小为目标进行配载优化。实践证明以牺牲食品品质为代价的低成本配载难以为继，如何在给定成本阈值内，提高生鲜食品品质成为关系民生及提升物流企业核心竞争力的关键难题，而“时间-温度”流数据是破解该难题的关键。本项目基于“逆优化”的学术思想，以调整配载车辆不同温区的温度设定值为切入点，在成本约束下最大化生鲜食品的配送品质。项目将基于“时间-温度”流数据的货架期定量预测方法引入配载模型，形成“时间-温度”流数据挖掘→成本最小化原模型→逆优化转化→最优化与智能化结合求解的一体化解决途径，并结合某大型生鲜物流企业开展应用研究。项目从食品安全视角优化生鲜配载计划，拟使配载计划由“成本”驱动向“品质”驱动转变，为成本控制下的生鲜食品品质提升提供了新的理论思路和方法，同时丰富和扩展了逆优化的研究。

项目针对生鲜农产品企业在物流环节中产品的品质与物流成本难以兼顾的难题，采用基于数据、 图像解析等方法对配载优化问题进行模型构建和算法设计等理论研究，以某浆果合作社、社区生鲜连锁店为背景进行了应用研究，并在上述研究基础上进行了一定的扩展研究。 . 针对生鲜农产品配载优化问题，考虑食品品质、不同食品存放限制等约束建立混合整数非线性规划模型，通过将货架期转化为食品品质等级，将模型线性化之后进行求解。针对生鲜农产品配载逆优化问题，在上述问题的配载方案基础上，以最小化温度设定调整量为优化目标建立逆优化模型，采用改进的遗传算法进行求解。此外将研究对象扩展到上游，对于给定的配载方案，针对不同货架期货品堆存顺序与装载顺序不一致的情况，为保证食品品质，以最小化装载过程的翻倒次数为目标建立整数线性规划模型，根据提炼的问题最优解性质，分别设计动态规划和列生成算法进行求解。对基于图像的等级标识样本进行训练，得到配载计划问题基于图像识别的模型参数。此外，以某社区生鲜连锁超市为背景开发决策支持系统。 . 项目将食品品质与物流成本在供应链环节中的矛盾以逆优化手段进行化解，所设计模型及算法为生鲜农产品配载决策过程提供了科学依据，提高了消费者对生鲜食品品质的满意度，同时对生鲜农产品冷链物流的高质量、可持续发展具有借鉴意义。