Students will understand the fundamental concepts and key algorithms of reinforcement learning (DP, MC, TD, Q-learning), and acquire knowledge of advanced methods such as DQN, Policy Gradient, and Actor–Critic. The course integrates lectures with hands-on practice, enabling students to implement and validate algorithms. Through this, they will develop foundational competencies to apply reinforcement learning to problem-solving in mechanical engineering and robotics.

Week 1 Course overview & Introduction to Deep Learning (basic concepts, perceptron, activation, loss)
Week 2 Deep Learning Review II – Deep Learning, Backpropagation, Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Applications in Mechanical Engineering
Week 3 History of Reinforcement Learning, Key Applications in Mechanical Engineering
Week 4 States, Actions, Rewards, Policies, Value Functions
Week 5 Policy Iteration and Value Iteration
Week 6 Monte Carlo (MC) Prediction and Control Basics
Week 7 SARSA and Q-Learning
Week 8 Exploration vs. Exploitation, ε-Greedy Policies
Week 9 Simple neural networks for RL – Value function approximation
Week 10 Deep Q-Network (DQN) / Midterm Exam
Week 11 Policy Gradient Methods
Week 12 Actor–Critic Methods
Week 13 Integration of RL and Deep Learning
Week 14 Applications and Case Studies
Week 15 Recent Trends and Advanced Topics in RL (e.g., A3C, DDPG, SAC)
Week 16 Final Summary, Project Presentations, and Course Wrap-up

The course will be delivered through lectures combined with hands-on practice.