讲座题目：《Neural combinatorial optimization model forlarge-scalevehicleroutingproblems withcomplexconstraints》

主讲嘉宾：张珂 博士

讲座主持：唐昕迪 助理教授

讲座时间：2025年6月11日（星期三）上午8:00--10:45

讲座地点：沙河校区主教407M

讲座摘要：The on-demand mobility service plays a crucial role in modern economies, where delivering reliable and responsive transport solutions carries substantial socioeconomic value. Conventional methods typically employ either mathematical formulations or heuristic techniques, yet frequently encounter difficulties in achieving trade-off between computational efficiency and solution accuracy. The emergence of neural combinatorial optimization has opened novel avenues for addressing these complex routing challenges. Through reinforcement learning paradigms, these methods can progressively assimilate operational insights from customer data patterns, enabling dynamic decision-making that reduces routing cost. However, key research gaps persist, including time constraint processing, vehicle coordination, and training stability. This report will show current progress in addressing these difficulties.

嘉宾简介：

张珂，清华大学博士，研究方向为大规模组合优化、智能交通系统、交通韧性管控。在Transportation Research Part C、IEEE TITS、IEEE TBD等领域知名期刊及会议发表论文20余篇，曾获美国TRB会议人工智能与先进计算应用委员会最佳博士论文奖，COTA会议最佳论文奖，清华大学优秀博士学位论文、国家奖学金等荣誉。主持省部级项目3项，以骨干身份参与多项国家重点研发计划项目。

主办单位：管理科学与工程学院