The movement of empty trucks (dead-heading) incurs significant costs and creates greenhouse gas emissions and road congestion. A strategy to tackle the dead-heading problem is to identify long-term backhaul collaboration opportunities, in which shippers and carriers generate frequent movement patterns (lanes) from historical truck movements in stage one, and then use the identified lanes as inputs in stage two for an optimization problem to create load-sharing contracts that eliminate empty backhauls. Where existing research has treated these as separate stages, we present an end-to-end integrated decision-making framework to connect these two stages and show that an integrated system would improve performance across diverse measures through improved lane generation and optimization. Our research offers a new design framework for backhaul collaboration that integrates spatial analytics methods into an optimization model, with improvements in financial, environmental, and social benefits. We use fine-grained GPS telematics data collected from two large logistics companies and evaluate potential lanes using spatial analytics techniques. Our framework delivers up to 75% potential improvements compared to a standard approach commonly used in practice.

中文翻译：

---

使用空间分析和优化的集成车道识别和长期回程协作的决策支持框架

空卡车的移动（空头行驶）会产生巨大的成本，并造成温室气体排放和道路拥堵。解决空头问题的策略是确定长期回程协作机会，其中托运人和承运人根据第一阶段的历史卡车运动生成频繁的运动模式（车道），然后使用已识别的车道作为第二阶段的输入解决优化问题以创建消除空回程的负载共享合同。现有研究将这些视为单独的阶段，我们提出了一个端到端集成决策框架来连接这两个阶段，并表明集成系统将通过改进车道生成和优化来提高各种措施的性能。我们的研究为回程协作提供了一个新的设计框架，将空间分析方法集成到优化模型中，从而改善财务、环境和社会效益。我们使用从两家大型物流公司收集的细粒度 GPS 远程信息处理数据，并使用空间分析技术评估潜在的车道。与实践中常用的标准方法相比，我们的框架可提供高达 75% 的潜在改进。