Date: November 30, 2020
Venue: Room 0212, Teaching Building 0#, Jiuli Campus
Lecturer: Professor Qin, Hu
About the Lecturer:
Professor Qin, Hu received his bachelor's and master's degrees from the School of Artificial Intelligence and Automation, Huazhong University of Science and Technology in 2002 and 2005, and got his Ph.D. from the Business School of City University of Hong Kong in 2011. His main research areas include Artificial Intelligence Optimization Algorithms, Network Planning, Vehicle Routing Optimization, Intelligent Logistics, Intelligent Warehousing, Intelligent Manufacturing, Data Mining and Machine Learning.
Professor Qin has completed two projects of National Natural Science Foundation of China and is currently presiding over one National Natural Science Foundation of China. He has published more than 40 papers in SCI/SSCI English journals, including 1 article on INFORMS Journal On Computing (IJOC), 3 articles in Transportation Science (TS) and 11 articles in European Journal of Operational Research (EJOR). On October 18, 2018, Professor Qin won the overall championship of JD Global Logistics Optimization Challenge (Intelligent Dispatching of Urban Logistics Transportation Vehicles). Professor Qin actively applies academic research results to corporate practice, creates value for the company, enhances its competitiveness, and has presided over and completed a number of corporate projects. His cooperative units include Huawei Technologies Co., Ltd., SF Express Co., Ltd., and Midea Group. , Blue Moon Co., Ltd., Guangzhou Procter & Gamble Co., Ltd., Shanghai Keyan Software Technology Co., Ltd., etc.
About the Lecture:
We study a general multiple depot vehicle scheduling problem in this paper. Given a fleet of vehicles based at multiple depots and a set of trips to be served, we aim to find a set of routes such that the total transportation cost is minimized and all trips are fulfilled. We call this problem “general” because many practical features have been considered, including heterogeneous vehicles, service start time windows, split loads, and the toll-by-weight scheme. These features, on the one hand, greatly increase the applicability of this problem; on the other hand, however, make the problem very challenging to solve. Based on the concept of cover, we formulate a set partitioning model for this problem. To solve the model, we propose two exact algorithms, namely, a branch-and-price algorithm and a branch and Benders decomposition algorithm. Both algorithms are built upon a branch-and-bound framework. Particularly, we design a column generation procedure equipped with an efficient label setting method to solve subproblems involved in algorithms. Computational experiments have been conducted, and the results have substantiated the effectiveness and efficiency of our approaches.