Learning Policy Sketches for Classical Planning
Title: Learning Policy Sketches for Classical Planning
SNIC Project: SNIC 2021/5-488
Project Type: SNIC Medium Compute
Principal Investigator: Hector Geffner <hector.geffner@liu.se>
Affiliation: Linköpings universitet
Duration: 2021-10-28 – 2022-11-01
Classification: 10201


In this work, we are trying to learn policy sketches for solving tractable classical planning domains in provably low polynomial time. A policy sketch R decomposes a problem into subproblems. The challenge is to automatically learn a suitable sketch that works on a large class of problems over a common domain. There are certain properties that a sketch has to satisfy such as being well-formed and has bounded and small sketch width in order to be suitable. We use answer set programming for finding suitable sketches automatically.