Course content

Introduction

Reinforcement learning (RL) is a powerful branch of machine learning that focuses on decision-making and sequential optimization. In recent years, RL has gained significant traction in finance and economics, where agents must learn to make optimal decisions under uncertainty and dynamic conditions. From algorithmic trading and portfolio management to macroeconomic modeling and policy design, RL offers a data-driven framework for solving complex problems.

This course introduces students to the fundamentals of RL and explores its applications in financial and economic contexts. Emphasis is placed on both theoretical foundations and practical implementation, including ethical considerations and interpretability.

Course Contents

This course will cover the following topics:

• Fundamentals of reinforcement learning: agents, environments, rewards, and policies
• Markov decision processes (MDPs) and dynamic programming
• Model-free methods: Q-learning, SARSA, and policy gradients
• Deep reinforcement learning: Deep Q-Networks, Actor critic methods, Proximal policy optimization
• Applications in finance: trading strategies, portfolio optimization, risk management
• Applications in economics: dynamic pricing, resource allocation, policy simulation, macroeconomic modeling
• Multi-agent reinforcement learning (MARL): coordination, competition, and decentralized decision-making
• Evaluation and interpretability of RL models
• Implementation using Python and libraries such as OpenAI Gym, Stable Baselines 3, TensorFlow, and PyTorch

Learning outcome

Knowledge

By the end of this course, the student shall be able to:

• Understand the core concepts and mathematical foundations of reinforcement learning
• Describe how RL algorithms operate and how they differ from supervised and unsupervised learning
• Explain how RL can be applied to solve problems in finance and economics
• Recognize challenges such as exploration vs. exploitation, sample efficiency, and stability
• Understand the principles of multi-agent reinforcement learning and its relevance to modeling strategic interactions in financial and economic systems

Skills

By the end of this course, the student shall be able to:

• Implement RL algorithms using Python and relevant libraries
• Design and evaluate RL models for financial and economic applications
• Apply RL to solve domain-specific problems within their master’s field
• Analyze and interpret the behavior of RL agents in dynamic environments
• Develop and test multi-agent RL systems for scenarios involving competition and cooperation

General competence

By the end of this course, the student shall be able to:

• Communicate the strengths and limitations of RL in finance and economics
• Reflect on ethical and practical implications of deploying RL systems
• Collaborate on projects involving machine learning and economic modeling

Learning methods and activities

Physical and digital lectures, self-study, project work, data exercises, and supervision by the lecturer or student assistants.

Compulsory assignments

• Mandatory presentation

Specific conditions

Recommended previous knowledge

Basic knowledge of programming (preferably Python) and machine learning is recommended. Students without this background are encouraged to engage in self-study using resources provided on Canvas.

Required previous knowledge

None

Course materials

Will be announced at the start of the semester.