基于关联网络的机器人移动货架系统储位分配优化方法

储位分配一直是仓储系统优化领域的热点和难点问题。近年来，机器人移动货架系统的出现给储位分配带来了新的、更大的挑战，使其不仅需要考虑商品的相关性，更应该关注各类商品在订单中的数量关系，这已超出了现有理论方法的适用范围。.针对机器人移动货架系统储位分配难题，本项目构建品类关系网络、复合关系网络、车辆路径网络这三层关联网络，通过网络元素与结构的层层映射，将该问题转化为一种新型车辆路径问题；建立包括商品选择与聚类、新型车辆路径问题建模与求解、储位方案还原的优化理论与方法；结合京东机器人移动货架系统的建设，开展应用研究，构建决策支持系统，提炼有价值的管理启示。.本项目为一类复杂决策问题创建了网络映射与问题转化的学术思想，使机器人移动货架系统储位分配可基于更加成熟的车辆路径问题的相关理论而加以解决，有利于增强优化理论面向复杂决策问题的适用性，提高移动货架系统储位分配的科学性、有效性、及智能性。

Storage allocation has always been a hot and difficult issue in the field of storage system optimization. In recent years, the emergence of the system with robot-moving racks brings new and greater challenges to storage allocation. Not only the relevance of goods, but also the quantitative relationship of various types of goods in orders should be considered, which is beyond the scope of the existing theoretical methods..For the problem of storage allocation in the system with robot-moving racks, this project constructs the three-layer associated networks of goods relationship network, composite relationship network, and vehicle routing network. By mapping network elements and structures among layers, the problem is transformed into a new type of vehicle routing problem. Optimization methodologies of selecting and clustering goods, modeling and solving the new type of vehicle routing problem, and restoring storage allocation scheme, are proposed. Along with the construction of the system with robot-moving racks in JingDong, the project carries out an applied research, develops a decision-making system, and extracts valuable managerial insights..This project puts forward a new academic idea of network mapping and problem transformation for a class of complex decision-making problems. This idea enables the storage allocation in the system with robot-moving racks to be resolved based on the more mature theories of vehicle routing problems. It is conducive to enhancing the applicability of optimization theory to complex decision-making problems, and improving the scientificity, effectiveness, and intelligence of the storage allocation in the system with mobile racks.