Adding Local Exploration to Greedy Best-First Search in Satisficing Planning

Authors: Fan Xie, Martin Müller, Robert Holte

AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments were run on a set of 2112 problems in 54 domains from the seven International Planning Competitions which are publicly available
Researcher Affiliation Academia Fan Xie and Martin Müller and Robert Holte Computing Science, University of Alberta Edmonton, Canada
Pseudocode Yes Algorithm 1 shows GBFS-LE.
Open Source Code No The paper does not provide an explicit statement or link for open-source code for the methodology described in this paper.
Open Datasets Yes Experiments were run on a set of 2112 problems in 54 domains from the seven International Planning Competitions which are publicly available
Dataset Splits No The paper does not explicitly provide training/test/validation dataset splits (e.g., percentages or absolute counts) for reproducing the experiments.
Hardware Specification Yes Experiments were run on a set of 2112 problems in 54 domains from the seven International Planning Competitions which are publicly available2, using one core of a 2.8 GHz machine with 4 GB memory and 30 minutes per instance.
Software Dependencies Yes All planners are implemented on the Fast Downward code base FD-2011 (Helmert 2006).
Experiment Setup Yes Parameters were set as follows: STALL_SIZE = 1000 for both algorithms. (MAX_LOCAL_TRY, LSSIZE) = (100, 1000) for GBFS-LS and (10, 100) for GBFS-LRW.