Monte Carlo Techniques in Planning

• Thesis title in Czech: Monte Carlo techniky v plánování
• Thesis title in English: Monte Carlo Techniques in Planning
• Key words: plánování, Monte Carlo Tree Search, učení HTN, logistické domény
• English key words: planning, Monte Carlo Tree Search, HTN-learning, transportation domain
• Academic year of topic announcement: 2017/2018
• Thesis type: rigorosum thesis
• Thesis language: angličtina
• Department: Department of Theoretical Computer Science and Mathematical Logic (32-KTIML)
• Supervisor: prof. RNDr. Roman Barták, Ph.D.
• Author: hidden - assigned and confirmed by the Study Dept.
• Date of registration: 09.08.2018
• Date of assignment: 09.08.2018
• Confirmed by Study dept. on: 09.08.2018
• Date and time of defence: 17.12.2018 00:00
• Date of electronic submission: 25.10.2018
• Date of submission of printed version: 25.10.2018
• Date of proceeded defence: 17.12.2018

Guidelines

The student will familiarize with existing Monte Carlo techniques such as Monte Carlo Tree Search and its derivatives and also with the classical formulation of planning problems. Based on this study, the student will propose how to exploit the Monte Carlo techniques to solve planning problems. This could be done by (semi-)automated reformulation of the planning problem or by proposing a set of modeling principles that would allow to manually formulate a planning problem to a form appropriate for the solving algorithm. A complementary approach is modifying the solving approach for the planning problems.

References

Cameron Browne, Edward Powley, Daniel Whitehouse, Simon Lucas, Peter I. Cowling, Philipp Rohlfshagen, Stephen Tavener, Diego Perez, Spyridon Samothrakis and Simon Colton: A Survey of Monte Carlo Tree Search Methods, IEEE Transaction on Computational Intelligence and AI in Games, Vol. 4, No. 1, March 2012

Malik Ghallab, Dana Nau, Paolo Traverso: Automated Planning: Theory and Practice, Morgan Kaufmann, 2004