Deep Reinforcement Learning for Wireless Communications and Networking: Theory, Applications and Implementation [Team-IRA]

✍ Scribed by Dinh Thai Hoang, Nguyen Van Huynh, Diep N. Nguyen, Ekram Hossain, Dusit Niyato

• Publisher: Wiley-IEEE Press
• Year: 2023
• Tongue: English
• Leaves: 280
• Edition: 1
• Category: Library

✦ Synopsis

Deep Reinforcement Learning for Wireless Communications and Networking

Comprehensive guide to Deep Reinforcement Learning (DRL) as applied to wireless communication systems

Deep Reinforcement Learning for Wireless Communications and Networking presents an overview of the development of DRL while providing fundamental knowledge about theories, formulation, design, learning models, algorithms and implementation of DRL together with a particular case study to practice. The book also covers diverse applications of DRL to address various problems in wireless networks, such as caching, offloading, resource sharing, and security. The authors discuss open issues by introducing some advanced DRL approaches to address emerging issues in wireless communications and networking.

Covering new advanced models of DRL, e.g., deep dueling architecture and generative adversarial networks, as well as emerging problems considered in wireless networks, e.g., ambient backscatter communication, intelligent reflecting surfaces and edge intelligence, this is the first comprehensive book studying applications of DRL for wireless networks that presents the state-of-the-art research in architecture, protocol, and application design.

Deep Reinforcement Learning for Wireless Communications and Networking covers specific topics such as:

• Deep reinforcement learning models, covering deep learning, deep reinforcement learning, and models of deep reinforcement learning
• Physical layer applications covering signal detection, decoding, and beamforming, power and rate control, and physical-layer security
• Medium access control (MAC) layer applications, covering resource allocation, channel access, and user/cell association
• Network layer applications, covering traffic routing, network classification, and network slicing

With comprehensive coverage of an exciting and noteworthy new technology, Deep Reinforcement Learning for Wireless Communications and Networking is an essential learning resource for researchers and communications engineers, along with developers and entrepreneurs in autonomous systems, who wish to harness this technology in practical applications.

✦ Table of Contents

ch1
ch1
1.1 Wireless Networks and Emerging Challenges
1.2 Machine Learning Techniques and Development of DRL
1.2.1 Machine Learning
1.2.2 Artificial Neural Network
1.2.3 Convolutional Neural Network
1.2.4 Recurrent Neural Network
1.2.5 Development of Deep Reinforcement Learning
1.3 Potentials and Applications of DRL
1.3.1 Benefits of DRL in Human Lives
1.3.2 Features and Advantages of DRL Techniques
1.3.3 Academic Research Activities
1.3.4 Applications of DRL Techniques
1.3.5 Applications of DRL Techniques in Wireless Networks
1.4 Structure of this Book and Target Readership
1.4.1 Motivations and Structure of this Book
1.4.2 Target Readership
1.5 Chapter Summary
References
ch2
2.1 Markov Decision Process
2.2 Partially Observable Markov Decision Process
2.3 Policy and Value Functions
2.4 Bellman Equations
2.5 Solutions of MDP Problems
2.5.1 Dynamic Programming
2.5.1.1 Policy Evaluation
2.5.1.2 Policy Improvement
2.5.1.3 Policy Iteration
2.5.2 Monte Carlo Sampling
2.6 Reinforcement Learning
2.7 Chapter Summary
References
ch3
3.1 Value‐Based DRL Methods
3.1.1 Deep Q‐Network
3.1.2 Double DQN
3.1.3 Prioritized Experience Replay
3.1.4 Dueling Network
3.2 Policy‐Gradient Methods
3.2.1 REINFORCE Algorithm
3.2.1.1 Policy Gradient Estimation
3.2.1.2 Reducing the Variance
3.2.1.3 Policy Gradient Theorem
3.2.2 Actor‐Critic Methods
3.2.3 Advantage of Actor‐Critic Methods
3.2.3.1 Advantage of Actor‐Critic (A2C)
3.2.3.2 Asynchronous Advantage Actor‐Critic (A3C)
3.2.3.3 Generalized Advantage Estimate (GAE)
3.3 Deterministic Policy Gradient (DPG)
3.3.1 Deterministic Policy Gradient Theorem
3.3.2 Deep Deterministic Policy Gradient (DDPG)
3.3.3 Distributed Distributional DDPG (D4PG)
3.4 Natural Gradients
3.4.1 Principle of Natural Gradients
3.4.2 Trust Region Policy Optimization (TRPO)
3.4.2.1 Trust Region
3.4.2.2 Sample‐Based Formulation
3.4.2.3 Practical Implementation
3.4.3 Proximal Policy Optimization (PPO)
3.5 Model‐Based RL
3.5.1 Vanilla Model‐Based RL
3.5.2 Robust Model‐Based RL: Model‐Ensemble TRPO (ME‐TRPO)
3.5.3 Adaptive Model‐Based RL: Model‐Based Meta‐Policy Optimization (MB‐MPO)
3.6 Chapter Summary
References
ch4
4.1 System Model and Problem Formulation
4.1.1 System Model and Assumptions
4.1.1.1 Jamming Model
4.1.1.2 System Operation
4.1.2 Problem Formulation
4.1.2.1 State Space
4.1.2.2 Action Space
4.1.2.3 Immediate Reward
4.1.2.4 Optimization Formulation
4.2 Implementation and Environment Settings
4.2.1 Install TensorFlow with Anaconda
4.2.2 Q‐Learning
4.2.2.1 Codes for the Environment
4.2.2.2 Codes for the Agent
4.2.3 Deep Q‐Learning
4.3 Simulation Results and Performance Analysis
4.4 Chapter Summary
References
ch5
ch5
5.1 Beamforming, Signal Detection, and Decoding
5.1.1 Beamforming
5.1.1.1 Beamforming Optimization Problem
5.1.1.2 DRL‐Based Beamforming
5.1.2 Signal Detection and Channel Estimation
5.1.2.1 Signal Detection and Channel Estimation Problem
5.1.2.2 RL‐Based Approaches
5.1.3 Channel Decoding
5.2 Power and Rate Control
5.2.1 Power and Rate Control Problem
5.2.2 DRL‐Based Power and Rate Control
5.3 Physical‐Layer Security
5.4 Chapter Summary
References
ch6
6.1 Resource Management and Optimization
6.2 Channel Access Control
6.2.1 DRL in the IEEE 802.11 MAC
6.2.2 MAC for Massive Access in IoT
6.2.3 MAC for 5G and B5G Cellular Systems
6.3 Heterogeneous MAC Protocols
6.4 Chapter Summary
References
ch7
7.1 Traffic Routing
7.2 Network Slicing
7.2.1 Network Slicing‐Based Architecture
7.2.2 Applications of DRL in Network Slicing
7.3 Network Intrusion Detection
7.3.1 Host‐Based IDS
7.3.2 Network‐Based IDS
7.4 Chapter Summary
References
ch8
8.1 Content Caching
8.1.1 QoS‐Aware Caching
8.1.2 Joint Caching and Transmission Control
8.1.3 Joint Caching, Networking, and Computation
8.2 Data and Computation Offloading
8.3 Data Processing and Analytics
8.3.1 Data Organization
8.3.1.1 Data Partitioning
8.3.1.2 Data Compression
8.3.2 Data Scheduling
8.3.3 Tuning of Data Processing Systems
8.3.4 Data Indexing
8.3.4.1 Database Index Selection
8.3.4.2 Index Structure Construction
8.3.5 Query Optimization
8.4 Chapter Summary
References
ch9
ch9
9.1 Adversarial Attacks on DRL
9.1.1 Attacks Perturbing the State space
9.1.1.1 Manipulation of Observations
9.1.1.2 Manipulation of Training Data
9.1.2 Attacks Perturbing the Reward Function
9.1.3 Attacks Perturbing the Action Space
9.2 Multiagent DRL in Dynamic Environments
9.2.1 Motivations
9.2.2 Multiagent Reinforcement Learning Models
9.2.2.1 Markov/Stochastic Games
9.2.2.2 Decentralized Partially Observable Markov Decision Process (DPOMDP)
9.2.3 Applications of Multiagent DRL in Wireless Networks
9.2.4 Challenges of Using Multiagent DRL in Wireless Networks
9.2.4.1 Nonstationarity Issue
9.2.4.2 Partial Observability Issue
9.3 Other Challenges
9.3.1 Inherent Problems of Using RL in Real‐Word Systems
9.3.1.1 Limited Learning Samples
9.3.1.2 System Delays
9.3.1.3 High‐Dimensional State and Action Spaces
9.3.1.4 System and Environment Constraints
9.3.1.5 Partial Observability and Nonstationarity
9.3.1.6 Multiobjective Reward Functions
9.3.2 Inherent Problems of DL and Beyond
9.3.2.1 Inherent Problems of DL
9.3.2.2 Challenges of DRL Beyond Deep Learning
9.3.3 Implementation of DL Models in Wireless Devices
9.4 Chapter Summary
References
ch10
10.1 DRL for Emerging Problems in Future Wireless Networks
10.1.1 Joint Radar and Data Communications
10.1.2 Ambient Backscatter Communications
10.1.3 Reconfigurable Intelligent Surface‐Aided Communications
10.1.4 Rate Splitting Communications
10.2 Advanced DRL Models
10.2.1 Deep Reinforcement Transfer Learning
10.2.1.1 Reward Shaping
10.2.1.2 Intertask Mapping
10.2.1.3 Learning from Demonstrations
10.2.1.4 Policy Transfer
10.2.1.5 Reusing Representations
10.2.2 Generative Adversarial Network (GAN) for DRL
10.2.3 Meta Reinforcement Learning
10.3 Chapter Summary
References
index