A stochastic routing service for electrical vehicles

• Main Author: Dong, Yubo
• Other Authors: Zhang Jie
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence; DRNTU::Engineering::Computer science and engineering::Computer systems organization::Computer system implementation
• Online Access: http://hdl.handle.net/10356/66673
• Institution: Nanyang Technological University

Electrical vehicles have limited range because of under developed battery technology. Routing an electrical vehicle is different from routing a traditional fuel vehicle as it needs to consider the remaining mileages in order to avoid running out of batteries. In this project, a Q-learning algorithm is used to address this need. The battery information is encoded together with location information to guide the vehicles. The project is divided into two phases. In the first phase, the routing algorithm was proposed. A simulator was implemented to tune the parameters and assess the performance of the proposed algorithm. The Q-learning algorithm solved the problem of finding a route that needs a few charging operations between the origin and destination. It showed much better performance when considering the remaining battery level or mileage of the vehicle. In the second phase, the algorithm was modified to solve the maximum probability of arrival problem in a stochastic environment. The algorithm provides a total solution that identifies the optimal directions to ensure maximum probability of arriving. It is able to differentiate the paths with less than 5% difference in probability. The Q-learning algorithm is successfully implemented as a routing service. It solved the problem of routing an electrical vehicle and the maximum probability of arrival problem, and showed great potential of becoming an intelligent navigation system. A few possible improvements are recommended as future works.