Dynamic Programming and Reinforcement Learning (A.Y. 2023/24)

Project work

• The project work can be carried out individually or in group (2 students per group)

• Each group must find 1 or 2 problems of interest and solve them

• The solution should include both DP and RL techniques

• The outcome of the project work is

  • a pdf file containing the problem formulation

  • a Jupyter Notebook containing the developed code and a discussion (with tables, graphs, plots, etc.) about

    • the aspects that have been analyzed

    • the obtained results

• The Jupyter Notebook must be turned in (via email) one week before the date of the oral exam (see below)

• Two weeks before the date of the oral exam (see below) each group must send me an email with the following information

  • name of the group members, together with their ID number (matricola)

  • description of the problems and the aspects that the students intend to investigate

Oral exam

• The oral exam includes a discussion about the project work and questions about the theory.

  • The discussion about the project work aims at verifying that students have correctly applied the studied methodologies and that they have correctly interpreted the obtained results. During the discussion, students may be required to write small pieces of code to investigate further aspects of the considered data sets.

  • Theoretical questions aim at verifying that students have acquired the basic notions of DP and RL.

Calendar of the Winter exam session

• Date of oral exam

  • February 7

  • February 28

• Deadline for turning in project work solution (one week before discussion)

  • January 31

  • February 21

• Deadline for turning in problems and project work description (one week before solution)

  • January 24

  • February 14