A Reinforcement Learning approach for lead vehicle routing in a semi-autonomous last-mile transportation system

02 June 2022, 12:00 
zoom & Room 206 
A Reinforcement Learning approach for lead vehicle routing in a semi-autonomous last-mile transportation system

Avital Shamir, M.Sc. student at the department of Industrial

2 June 2022, 12:00 PM, Room 206& via zoom


In semi-autonomous transportation systems, vehicle autonomy capabilities are utilized in a partial manner, so as to adhere to current regulations. The multi-layered personal transit system is a special design in which convoys composed of human-driven lead vehicles and autonomous trailers provide station-to-station transportation. At the proximity of stations, trailers can detach from the convoy and travel autonomously to enter/exit the stations. In a previous study, the assignment of passengers to trailers and the routing of the trailers was determined dynamically, while static circular routes were determined for the lead vehicles. In this study, we introduce for each circular lead vehicle route a potential shortcut, with the aim of determining dynamically when a lead vehicle should take shortcuts. We present an abstraction of the system and formulate the resulting decision problem as a Markov Decision Process in which the state represents the open trailer tasks, an action represents the shortcut decisions, and the penalty represents the number of trailers waiting to be served. We develop an event-based simulation framework to represent the system’s dynamics and employ a Reinforcement Learning approach to dynamically decide upon the lead vehicle shortcuts. The obtained policies are shown to outperform fixed lead vehicle route plans and simple-rule based dynamic shortcut policies.


Avital Shamir,  
is a research student in the department of Industrial Engineering at Tel-Aviv University. She holds a B.Sc. degree in Industrial Engineering and Management and a M.Sc. degree in Industrial Engineering from Ben Gurion University with a specialization in Data Science and Business Analytics. Her research focuses on reinforcement learning approaches to optimizing semi-autonomous transportation systems. For the last ten years, Avital has been a research scientist at Intel Corporation. The research is supervised by Dr. Mor Kaspi.

Tel Aviv University makes every effort to respect copyright. If you own copyright to the content contained
here and / or the use of such content is in your opinion infringing, Contact us as soon as possible >>