报告时间：2023年4月24日 10:00开始
报 告 人：唐与聪（南京航空航天大学 讲师）
报告地点：9-113
报告题目：ADMM-based Methods for a Robust Aircraft Rerouting Model with Slot Availability, Airport Capacity and Maintenance Flexibility Constraints
报告摘要：Once a disruption occurs, airline decision makers must give a recovery plan in a very short time. Otherwise, disruptions may interrupt the airline operation and lead to great economic and reputation losses. In this paper, we address the aircraft rerouting problem (ARP) in the
operation level. We develop a realistic aircraft rerouting model (ARM) with slot availability, airport capacity constraints and maintenance exibility simultaneously and explicitly. Besides, we consider reducing the critical connections which are tight in time and more susceptible to disruptions in order to increase the robustness of the recovery plan. When the number of ights is large, there are many variables and constraints in the ARM and thus it is quite difficult to solve the
ARM.We propose two methods based on alternating direction method of multipliers (ADMM) and column generation (CG), namely the Linearized ADMM-CG and the Hybrid ADMM-CG. Specifically, a large ARP can be transformed into multiple subproblems with ADMM. To further reduce the solving difficulty, we utilize the problem properties and the proximal-term technique to linearize the ADMM subproblems. Then CG is used to solve the linearized subproblems efficiently. Test cases show that our methods can recover aircraft routes effectively in a reasonable time.
报告人简介： 唐与聪，博士毕业于中国科学院大学。研究领域：图论及其应用，航空运输优化，机器学习。目前担任中国工业与应用数学学会通信的数学基础专委会副秘书长。在《Journal of Graph Theory》、《European Jounal of Combinatorics》、《Discrete Mathematics》等杂志发表论文10余篇。主持国家自然科学基金一项，参与国家自然科学基金重点项目一项。
