Subject Datasheet

Download PDF
Budapest University of Technology and Economics
Faculty of Transportation Engineering and Vehicle Engineering
1. Subject name Reinforcement Learning for vehicle control
2. Subject name in Hungarian Megerősítéses tanulás a járműirányításban
3. Code BMEKOKAD017 4. Evaluation type exam grade 5. Credits 3
6. Weekly contact hours 2 (0) Lecture 0 (0) Practice 0 (0) Lab
7. Curriculum
PhD Programme
 8. Role
Specific course
9. Working hours for fulfilling the requirements of the subject 90
Contact hours 28 Preparation for seminars 14 Homework 30
Reading written materials 0 Midterm preparation 0 Exam preparation 18
10. Department Department of Control for Transportation and Vehicle Systems
11. Responsible lecturer Dr. Bécsi Tamás
12. Lecturers Dr Bécsi Tamás, Dr. Aradi Szilárd
13. Prerequisites  
14. Description of lectures
Problem solving, placement in machine learning. Heuristics, dynamic and static heuristics. Effectiveness and complexity of algorithms. Curse of dimensions. The Markov decision model, the hidden Markov decision model. Traceability problem. Classic solutions for self-learning systems, case study for routing algorithms. Fundamentals of neural networks, supervised teaching, general network structures. Discrete, continuous and regular tasks. Reverse learning, Imitation learning. Demonstrator and demonstration, policy, loss function and algorithms. Value based learning, Q-learning. The exploration-exploitation dilemma. Variations of Q learning, Deep Q, DQN. Behavior based learning algorithms, Policy gradients, deterministic, and stochastic policy.
15. Description of practices
 
16. Description of labortory practices
 
17. Learning outcomes
A. Knowledge   B. Skills   C. Attitudes   D. Autonomy and Responsibility
18. Requirements, way to determine a grade (obtain a signature)
Final exam and three homeworks.
19. Opportunity for repeat/retake and delayed completion
 
20. Learning materials
 
Effective date 27 November 2019 This Subject Datasheet is valid for Inactive courses