SPEAKER: 李纪柳 教授, 西北工业大学

TIME/DATE: 2026年5月13日10：00

CLASSROOM: 同济大厦A楼309教室

ABSTRACT：

Real-world applications often require tracking resource usage over a planning horizon. The Worker Team Formation and Routing Problem (WTFRP) is one such application, which focuses on optimizing airport ground-handling operations. The WTFRP involves determining trips with time windows, forming teams with hierarchical and downgrade-compatible skill levels, and accounting for task execution times tied to team composition. We must track worker counts for each skill level throughout the planning horizon. In contrast to the most common approach of dividing the planning horizon into small intervals (discretization), we introduce an exact solution method for the WTFRP directly in the original continuous-time space (without discretization). We begin by formulating the WTFRP as an infinite integer programming model with an infinite number of variables and constraints. To tackle the linear programming (LP) relaxation of this formulation, we propose a novel approach called Effective Time-Point Detection (EFFECTION). This method dynamically identifies time-points to simultaneously generate columns and constraints, addressing the challenge of infinite variables and constraints. Building on EFFECTION, we develop a method termed Enumeration-Check-and-Switch (ECS). Enumeration uses variable-fixing techniques to identify promising variables. Check introduces a verification procedure to ensure feasibility. Switch is a model-switching mechanism alternating between the original formulation and a relaxed version. Additionally, the idea of EFFECTION is integrated into the process for generating the necessary variables and constraints. In the computational experiments, we validate the efficacy of each algorithmic feature. On a set of known instances, we show that ECS not only produces better solutions but also achieves them significantly faster—up to two orders of magnitude—compared to the existing method. Furthermore, the results on a newly generated set of instances confirm the robustness of ECS across time granularities and underscore the importance of considering finer time resolutions, which leads to higher-quality solutions for airport ground-handling operations.

GUEST BIO：
李纪柳，博士，西北工业大学教授、博士生导师，陕西省青年高层次引进人才。他毕业于华中科技大学并获博士学位，曾赴香港理工大学开展访问交流工作。李纪柳教授擅长精确算法、强化学习等智能决策方法，长期致力于复杂物流网络中重要优化问题的研究。2023年，他荣获管理科学与工程学会“优秀博士论文奖”；2026年，荣获陕西省第十七次哲学社会科学优秀成果一等奖。他主持承担国家自然科学基金青年项目（C类）及面上项目，研究成果先后发表于IJOC（UTD 24）、Transportation Research Part B/E、EJOR、Omega等国际顶级及重要期刊。同时，他担任陕西运筹学会常务理事，兼任IJOC、POM、EJOR、Transportation Research Part B/E/C、Omega等国际期刊匿名审稿人。