时变交通网络下的大规模物流运输时空过程建模及优化

Rapid growing logistics industry expects large scale, highly dynamic and high quality transportation service, which is significantly affected by the varying urban traffic. Modeling and optimization for logistics in time varying transportation network is a great challenge. Due to the fast rising computational complexity, efficient solving for logistics transportation is very difficult. Form the spatial-temporal perspective, this project investigates the modeling and optimization for large scale time varying logistics transportation problem (TVLTP). A spatio-temporal process model is proposed to finely represent complex logistics lifecycle. A multi-objectives optimization model is designed to balance between logistics cost and service level. To address the computational complexity issue, dimensionality reduction and speed-up strategies based on the spatial-temporal neighborhood are proposed. A spatial-temporal optimization framework is developed to efficiently solve TVLTP. For the real-time demand case, its spatial-temporal affected solution space is found and a dynamical optimization mechanism is designed to provide fast response. Finally, to demonstrate the applicability of proposed models and algorithms, a prototype logistics transportation service system (LTSS) is developed using real-world spatial-temporal transportation data and logistics instances. The development of fine modeling and efficient optimizing will provides fast preplanning and adjustment for large scale city logistics cases. It is expected that the LTSS not only promotes logistics efficiency, but also contributes to the intelligent logistics management and operation.

蓬勃发展的物流业需要大规模、高动态、高质量的物流运输服务。城市交通状态具有时空变化特性，对物流运输有重要影响。时变交通网络下的物流运输建模与优化是重要科学问题。物流运输优化问题的计算复杂度随规模扩大急剧增长，快速优化非常困难。本课题从时空视角，研究大规模物流运输的时空过程精细化建模方法；提出物流总费用和服务水平均衡的多目标优化模型。在此基础上，研究基于时空邻近性的高效率物流运输时空过程启发式优化方法，提出时空降维策略，减少问题复杂性；研究基于局部搜索的加速优化策略，提高优化效率。顾及实时客户需求，研究其时空影响区域发现策略，提出物流时空过程动态优化方法。最后，对模型和算法进行综合实验验证。解决了复杂交通环境下物流运输时空过程精细化表达与高效率优化难题，在短时间内提供高质量的物流运输方案及动态调整，提高物流效率，降低物流费用，为智慧化物流管理与运营提供理论指导与技术支撑。

不断增长的物流运输需求需要大规模、高动态、高质量的物流运输服务。城市交通状态具有时空变化特性，对物流运输有重要影响。时变交通网络下的物流运输建模与优化是重要科学问题。物流运输优化问题的计算复杂度随规模扩大急剧增长，快速优化非常困难。针对时变交通网络下的物流运输，本课题研究了大规模物流运输的时空过程模型；提出面向物流运输的多目标优化模型，研究多仓库物流车辆路径优化，基于时空邻近性的高效率物流运输时空过程启发式优化和空间并行优化、动态优化等系列方法，解决了城市物流优化的计算复杂性和动态性；对模型和算法进行综合实验验证。解决了复杂交通环境下物流运输时空过程精细化表达与高效率优化难题，在短时间内提供高质量的物流运输方案及动态调整，提高物流效率，降低物流费用。研究成果在深圳市交通运输委员会、广州奥格智能科技有限公司得到了应用，为智慧化物流管理与运营提供理论指导与技术支撑。