Deep Reinforcement Learning for Crowdsourced Urban Delivery: System States Characterization, Heuristics-guided Action Choice, and Rule-Interposing Integration
This paper investigates the problem of assigning shipping requests to ad...
Deep reinforcement learning (DRL) is an emerging methodology that is tra...
Tanvir Ahamedverfied profile
Currently pursuing my PhD in Transportation from University of Illinois at Chicago and working as a Research Assistant in a National Science Foundation (NSF) funded project [CMMI #1663411].
Last-mile urban delivery has been experiencing remarkable research attention in last five years due to explosive growth of e-commerce worldwide. As e-commerce is expanding, traditional employee and vehicle asset-based deliveries are expected to augment many negative consequences. In addition, traditional means for delivery may find the increasingly frequent (on-demand) delivery requirement more difficult, especially for food, grocery and retail.
I investigate the possibility of efficiently using ordinary people (i.e., "crowdsourcees") who can walk, bike or drive during transporting freight within cities. My work focuses on developing efficient heuristics for crowdsourcees-shipping requests assignment while proactively tackling the spatial imbalance of crowdsourcees in the system. Another recent work considers deep reinforcement learning for solving the same problem and justifying rich implication of artificial intelligence for vehicle routing problem.