Symmetry Pruning in Domain Descriptions

Abstract:

In the process of grounding domain descriptions of planning tasks, symmetries in operator descriptions may lead to multiple operators which don't differ in their semantics. If the domain is not carefully designed, this may lead to thousands of copies of the same operator. We adapt symmetry reduction techniques from state-based search and deploy these techniques before the process of grounding is initiated. Results from novel IPC domains indicate that performing symmetry reduction on the domain description level reduces the number of final operators by several orders of magnitude.

 

Biography

Florian is a PhD student in the Research Group on the Foundations of Artificial Intelligence at the University of Freiburg.

Florian Geißer recived a Master of Science in computer science from the University of Freiburg in January 2014. Since February 2014 he has been working in the Research Group on the foundation of Artificial Intelligence headed by Prof. Dr. Bernhard Nebel.

Date & time

2–3pm 31 May 2018

Location

Room:Systems Area Level 2

Speakers

Florian Geißer (from University of Freiburg)

Contacts

Updated:  1 June 2019/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing