Planning for Artificial Intelligence

The course cannot be taken simultaneously with:

Artificial Intelligence Planning (B4M36PUI)

The course is a substitute for:

Artificial Intelligence Planning (B4M36PUI)

Garant předmětu:

Michal Pěchouček

Lecturer:

Stefan Edelkamp, Antonín Komenda, Jan Mrkos, Michaela Urbanovská

Tutor:

Stefan Edelkamp, Antonín Komenda, Jan Mrkos, Michaela Urbanovská

Supervisor:

Department of Computer Science

Synopsis:

The course covers the problematic of automated planning in artificial intelligence and focuses especially on domain independent models of planning problems: planning as a search in the space of states (state-space planning), in the space of plans (plan-space planning), heuristic planning, planning in graph representation of planning problems (graph-plan) or hierarchical planning. The students will also learn about the problematic of planning under uncertainty and the planning model as a decision-making in MDP and POMDP.

Requirements:

Syllabus of lectures:

1. Introduction to the problematic of automated planning in artificial intelligence

2. Representation in form of search in the space of states (state-space planning)

3. Heuristic planning using relaxations

4. Heuristic planning using abstractions

5. Structural heuristics

6. The Graphplan algorithm

7. Compilation of planning problems

8. Representation of the planning problem in form of search in the space of plans (plan-space planning)

9. Hierarchical planning

10. Planning under uncertainty

11. Model of a planning problem as a Markov Decision Process (MDP)

12. Model of a planning problem as a Partially Observable Markov Decision Process (POMDP)

13. Introduction to planning in robotics

14. Applications of automated planning

Syllabus of tutorials:

1. Planning basics, representation, PDDL and planners

2. State-space planning, Assignment 1

3. Relaxation heuristics, Assignment 1 Consultations

4. Abstraction heuristics, Assignment 1 Deadline

5. Landmark heuristics, Assignment 1 Results/0-point Deadline

6. Linear Program formulation of heuristics

7. Compilations

8. Partial-order planning

9. Hierarchical Planning

10. Planning with uncertainty, Assignment 2

11. Planning for MDPs, Assignment 2 Consultations

12. Planning for POMDPs, Assignment 2 Consultations

13. Monte Carlo tree search, Assignment 2 Deadline

14. Consultations of exam topics, Assignment 2 Results/0-point Deadline, Credit

Study Objective:

Study materials:

* Malik Ghallab, Dana Nau, Paolo Traverso: Automated Planning: Theory & Practice, Elsevier, May 21, 2004

* https://www.coursera.org/course/aiplan

Note:

Further information:

https://cw.fel.cvut.cz/b192/courses/be4m36pui/start

Time-table for winter semester 2022/2023:

Time-table is not available yet

Time-table for summer semester 2022/2023:

	06:00–08:0008:00–10:0010:00–12:0012:00–14:0014:00–16:0016:00–18:0018:00–20:0020:00–22:0022:00–24:00
Mon	roomKN:E-301 Chrpa L. Edelkamp S. 09:15–10:45 (lecture parallel1) Karlovo nám. Šrámkova posluchárna K9roomKN:E-311 Urbanovská M. Mrkos J. 12:45–14:15 (lecture parallel1 parallel nr.101) Karlovo nám. Lab K311
Tue
Wed
Thu
Fri

The course is a part of the following study plans:

• Open Informatics - Artificial Intelligence (compulsory course of the specialization)