Topic：Distributed Computing Approaches for Large-Scale Traffic Assignment Problem

Lecturer：Professor Liu Zhiyuan ---- Southeast University

Moderator：Professor Xiao Feng ---- School of Business Administration

Time：NOV. 8^th, 2020. 15:00-16:00(GMT+8)

Live Broadcast Via Tencent MeetingID:564 165 641

(International Platform: VOOV Meeting ID: 564 165 641)

Host: School of Business Administration, SWUFE Scientific and Research Department

About the Lecturer:

Liu Zhiyuan, Professor, Ph.D supervisor and Vice President of School of Transportation, Southeast University, Head of Complex Traffic Network Research Center, Ph.D supervisor of School of Cyberspace Security of Southeast University, and visiting professor of Monash University of Australia. He was selected as the excellent youth of the National Science Foundation of China, the "Entrepreneurship and Innovation Talents" of Jiangsu Province, the "Youth Entrepreneurship and Innovation Talents" of Jiangsu Province, and the "Young Chief Professor" of Southeast University, and was awarded the "May Fourth Youth Medal" of Southeast University. In 2011, he graduated from the National University of Singapore with a doctor's degree, and then stayed in the University for postdoctoral research for one year. Since 2015, he returned to work in the School of Transportation of Southeast University. Before returning to China, he worked in the Department of Civil Engineering of Monash University in Australia as a lecturer and Ph.D supervisor. From December 2017 to January 2018, he was a visiting scholar in Department of mathematics, University of Melbourne, Australia. The main research fields include planning and management of transportation network, analysis and modeling on transportation big data , public transportation, multi-mode logistics network, etc. So far, more than 100 academic papers have been published in these fields, and more than 70 papers have been retrieved by SCI / SSCI journals and been cited by Google Scholar for more than 2000 times. He served as associate editor of ASCE Journal of Transportation Engineering and IET Intelligent Transport Systems, as well as the editorial board member of International Journals, such as Transportation Research Part E (SCI / SSCI), Transportation Research Record (SCI), Journal of Transport And Land Use (SSCI). He has also guided students to win a number of domestic and foreign big data algorithm Competition Awards (all top three), including the KDD CCUP Champion (2020), and the second place (2019) in the "Big Data Competition World Cup", and other IJCAI (Champion, 2019), which are also the top three international competitions of artificial intelligence. In addition, it also includes the champion of the algorithm challenge competition of Alibaba Tianchi competition in 2016, the runner up of the first didi algorithm competition in 2016, the excellent paper award of the data analysis competition of the United States TRB conference in 2017, the second runner up of the 2017 CCF big data and computing intelligence competition, and the 2017 UCAR artificial intelligence cup champion (IEEE Computer) In addition, it also won the first prize in big data competition of Digital China innovation competition in 2019.

About the Lecture:

Traffic assignment is a fundamental tool to evaluate the flow distribution pattern in a transport network. As one of the most recognized theory for traffic assignment, user equilibrium is widely investigated and implemented. Most of the existing algorithms for the user equilibrium-based traffic assignment problem are developed and implemented sequentially. This study aims to study and investigate the parallel computing approach to utilize the widely available parallel computing resources. The Parallel Block-Coordinate Descent (PBCD) algorithm is developed based on the path-based algorithm, i.e, the improved path-based gradient projection algorithm (iGP). A parallel block-coordinate method is proposed to replace the widely used Gauss-Seidel method for the procedure of path flow adjustment. To further improve the robustness/performance of the algorithm, the iPBCD algorithm is developed based on the state-of-the-art PBCD algorithm. A different type of flow update policy is studied and investigated intensively. The block size is determined using a sensitive test, and five indices-grouping rules are compared. Besides, a greedy update order of block indexes is introduced to compare with the cyclic scheme. Moreover, a new algorithm is developed based on the alternating direction method of multipliers (ADMM). In order to take use of the ADMM, the network links should be grouped into several blocks, where the links in the same block are disconnected. This link grouping problem falls into the category of edge-coloring problem in graph theory, and it follows the Vizing theorem. Numerical examples show that the proposed algorithms perform well in convergence and efficiency and can significantly reduce the computing time.