Designing a dynamic path guidance system based on electronic maps by using Q-learning

Liang Zou; Jianmin Xu; Lingxiang Zhu

doi:10.1117/12.658569

4 January 2006 Designing a dynamic path guidance system based on electronic maps by using Q-learning

Liang Zou, Jianmin Xu, Lingxiang Zhu

Proceedings Volume 5985, International Conference on Space Information Technology; 59855A (2006) https://doi.org/10.1117/12.658569
Event: International Conference on Space information Technology, 2005, Wuhan, China

Abstract

Shortest path problem from one origin node to one destination node in non-FIFO (First In First Out) dynamic networks is an unsolved hard problem in dynamic path guidance system. A new approach based on Q-learning is adopted to solve the problem based on electronic maps in this paper. The approach uses geographical information on electronic maps to define Q-learning's value function. Q-learning algorithm's strategy train learning method and training process on path searching are presented. Finally based on Guangzhou City's electronic map, we randomly generate a dynamic network containing 20000 nodes, 40000 links and 144 time intervals, which do not satisfy FIFO to test the approach proposed in this paper. The approach is implemented with this dynamic network and its computational performance is analyzed experimentally. The experimental results prove the effectiveness of the approach.

Citation Download Citation

Liang Zou, Jianmin Xu, and Lingxiang Zhu "Designing a dynamic path guidance system based on electronic maps by using Q-learning", Proc. SPIE 5985, International Conference on Space Information Technology, 59855A (4 January 2006); https://doi.org/10.1117/12.658569

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available