Unit 1

Introduction to Reinforcement Learning – History of Reinforcement Learning – Elements of Reinforcement Learning – Limitations and scope.

Unit 2

Multi-armed Bandits – Finite Markov Decision Processes – Dynamic Programming – Policy evaluation – Policy improvement – Policy Iteration – Value Iteration.

Unit 3

Monte Carlo Methods – Monte Carlo prediction – Monte Carlo control – Incremental Implementation – Temporal- Difference Learning – TD prediction – Q-Learning – n-step Bootstrapping.

Unit 4

Planning and Learning with Tabular Methods – Models and planning – Prioritized sweeping – Trajectory sampling – Heuristic search – Rollout algorithms.

Course Objectives:

• To provide a solid introduction to the field of reinforcement learning.
• To enable the students to learn about the core challenges and approaches, including exploration and exploitation.
• To expose the students to techniques like Monte Carlo and tabular methods.

Outcomes Course:

After completing this course, students should be able to
CO1: Demonstrate sound understanding of the foundations of Reinforcement Learning
CO2: Demonstrate proficiency in Multi-armed Bandits and Markov Decision Processes
CO3: Apply Monte Carlo Methods and Temporal-Difference Learning
CO4: Apply Tabular Methods in Planning and Learning
CO5: Employ Reinforcement Learning Concepts in Real-world Applications

CO-PO Mapping

1. Richard S. Sutton and Andrew G.Barto, Reinforcement Learning, MIT Press, Second Edition, 2018.