Development of a Neural Network Based Q Learning Algorithm for Traffic Signal Control

Li Bi Fu; Kil To Chong

doi:10.4028/www.scientific.net/AEF.2-3.91

Paper Titles

Synchronic Fuzzy Control of Master and Slave Arc in Twin-Wire Pulsed Metal Active Gas Welding
p.69

Control of Pushing and Pulling Motion of a Wheeled Inverted Pendulum Based on a Reduced Order Observer
p.74

A Sun Tracking PV System Equipped Greenhouse Control System for Educational Purpose
p.80

Output Feedback Control of a Ball and Beam System Based on Jacobian Linearization under Sensor Noise
p.85

Development of a Neural Network Based Q Learning Algorithm for Traffic Signal Control
p.91

Multi-Channel Active Noise Control System Designs with Fuzzy Logic Stabilized Algorithms
p.96

Removal of Ocular Artifact from EEG Using Constrained ICA
p.105

HHT Based Analysis on Experiments of Rotor System with Rub-Impact
p.111

Energy Analysis Method Based on Wavelet Transform for Sensor Fault Diagnosis
p.117

HomeAdvanced Engineering ForumAdvanced Engineering Forum Vols. 2-3Development of a Neural Network Based Q Learning...

Development of a Neural Network Based Q Learning Algorithm for Traffic Signal Control

Abstract:

As one kind of reinforcement learning method, Q learning algorithm has already been proved to achieve many significant results in traffic signal control area. However, when the state of Markov Decision Process is very big or continuous, the computation load and the memory load will become very big and can not be solved then. Therefore, this paper proposed a neural network based Q learning algorithm to solve this problem known as “Curse of Dimensionality”. This new method realized generalization of conventional Q learnig algorithm in huge and continuous state space as neural network is a very effective value function approximator. Experiment has been implemented upon an isolated intersection and simulation results show that the proposed method can improve the traffic efficiency significantly than the conventional Q learning algorithm.

By email View Pdf

You have full access to the following eBook

Read eBook

Info:

Periodical:

Advanced Engineering Forum (Volumes 2-3)

Pages:

91-95

DOI:

https://doi.org/10.4028/www.scientific.net/AEF.2-3.91

Citation:

Cite this paper

Online since:

December 2011

Authors:

Li Bi Fu, Kil To Chong

Keywords:

Neural Network (NN), Q Learning

Export:

RIS, BibTeX

Permissions:

Creative Commons CC BY 4.0

Citation:

References

[1] R. Sutton. Learning to predict by the methods of temporal difference. Machine Learning, 1988, 3: pp.9-44.

Google Scholar

[2] Lu Shoufeng, Liu Ximin and Dai Shiqiang. Q learning for adaptive traffic signal control based on delay minimization strategy. IEEE International Conference on Networking Sensing and Control, 2008, pp.687-691.

DOI: 10.1109/icnsc.2008.4525304

Google Scholar

[3] Wei Wu, Gong Shufeng and Liu Hongxiu. A coordinated urban traffic signal control approach based on multi-agent. INES'09 Proceedings of the IEEE 13th international conference on Intelligent Engineering Systems , (2009).

DOI: 10.1109/ines.2009.4924773

Google Scholar

[4] Marco Wiring. Multi-agent reinforcement learning for traffic light control. ICML '00 Proceedings of the Seventeenth International Conference on Machine Learning, (2000).

Google Scholar

[5] Mark Humphrys. Action selection methods using reinforcement learning. A dissertation submitted for the degree of Doctor of Philosophy in the University of Cambridge, (1997).

Google Scholar