声明
Abstract
Content
Chapter 1 Introduction
1.1 Background and significance of subject research
1.2 State of the art in the country and outside
1.2.1 State of the art in China
1.2.1 State of the art outside the country
1.2.3 Research objective and content
1.3 Organizational structure of this thesis
Chapter 2 Literature Review
2.1 Introduction
2.2 Traffic congestion related knowledge
2.2.1 Types of traffic congestion
2.2.2 Evaluation Index of Traffic Congestion
2.3 Tools and libraries explanation
2.3.1 Python
2.3.2 TensorFlow
2.3.3 NumPy
2.3.4 TKinter
2.4 Markov Decision Process
2.5 Deep Reinforcement Learning
2.5.1 Deep Learning
2.5.2 Reinforcement Learning
2.5.3 Deep Reinforcement Learning
2.5.4 Multiagent environment for Deep Reinforcement Learning
2.6 Summary of the chapter
Chapter 3 Deep Reinforcement leaming route traffic guidance approach
3.1 Introduction
3.2 Environment Setup and parameters
3.3 Multiagent Deep Reinforcement Learning
3.3.1 Reinforcement Learning analysis
3.3.2 The Deep Q-Network
3.3.3 Multiagent analysis
3.3.4 Kinematic constraints
3.4 Summary of the chapter
Chapter 4 Results and discussion
4.1 Introduction
4.2 Traffic as a space-time event
4.3 Deep Reinforcement Learning
4.3.1 Simple case analysis
4.3.2 Results of two agents and two destinations
4.3.3 Results of seven agents and seven destinations
4.3.4 Results of seven agents and seven destinations random positions
4.3.5 Results of twenty agents and twenty destinations random positions
4.3.6 Result of eighty agents and eighty destinations random positions
4.4 Discussions
Chapter 5 Conclusions
5.1 Thesis Summary
5.2 Short Term Future W6rk
REFERENCES
PUBLICATIONS DURING THE MASTER DEGREE
ACKNOWLEDGMENTS
沈阳理工大学;