Title:	Learning Dynamic Algorithms for Automated Planning
DNr:	NAISS 2024/23-476
Project Type:	NAISS Small Storage
Principal Investigator:	Jendrik Seipp <jendrik.seipp@liu.se>
Affiliation:	Linköpings universitet
Duration:	2024-09-01 – 2025-09-01
Classification:	10201
Homepage:	https://mrlab.ai/projects/
Keywords:

Abstract

We are applying for "NAISS Small Storage" for the continuation project “Learning Dynamic Algorithms for Automated Planning" (NAISS 2024/5-421). As our project has progressed successfully, we have started to produce more data to better learn dynamic algorithms for automated planning. Unfortunately, we are now frequently running out of storage space. In particular, the current limitation of one million files is a burden on the successful continuation of the project. The main reason for this is that our experiments with modern planners generate numerous files for running and evaluating the experiments. In addition, we are currently working with planning software that can generate not just a single solution, but multiple (sometimes thousands of different) solutions for a single planning task. We plan to use such solvers to generate a dataset of solutions from which we can learn. This requires that we have more storage space available to generate the data. We hope and are convinced that increasing the memory for our project will have a direct impact on the success and speed of our project, as it will help us to run larger and more experiments.