Autonomous-Vehicle Public Transportation System: Scheduling and Admission Control

• Published in: IEEE transactions on intelligent transportation systems Vol. 17; no. 5; pp. 1210 - 1226
• Main Authors: Lam, Albert Y. S., Yiu-Wing Leung, Xiaowen Chu
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.05.2016; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Admission control; Autonomous vehicle (AV); bilevel optimization; Car sharing; Flexibility; Mathematical analysis; Production scheduling; Public transportation; Quality of service; Roads; Scheduling; smart city; Transportation; Transportation services; Vehicles; Vehicular ad hoc networks; admission control; car sharing; bilevel optimization; Autonomous vehicle (AV); smart city
• ISSN: 1524-9050; 1558-0016
• DOI: 10.1109/TITS.2015.2513071

Technology of autonomous vehicles (AVs) is becoming mature, and many AVs will appear on roads in the near future. AVs become connected with the support of various vehicular communication technologies, and they possess a high degree of control to respond to instantaneous situations cooperatively with high efficiency and flexibility. In this paper, we propose a new public transportation system based on AVs. It manages a fleet of AVs to accommodate transportation requests, offering point-to-point services with ride sharing. We focus on the two major problems of the system: scheduling and admission control. The former is to configure the most economical schedules and routes for the AVs to satisfy the admissible requests, whereas the latter is to determine the set of admissible requests among all requests to produce maximum profit. The scheduling problem is formulated as a mixed-integer linear program, and the admission control problem is cast as a bilevel optimization, which embeds the scheduling problem as the major constraint. By utilizing the analytical properties of the problem, we develop an effective genetic-algorithm-based method to tackle the admission control problem. We validate the performance of the algorithm with real-world transportation service data.