Revisiting Dominance Pruning in Decoupled Search

Authors: Daniel Gnad11809-11817

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show empirically that all our improvements are indeed beneficial in many standard benchmarks.
Researcher Affiliation Academia Daniel Gnad Saarland University Saarland Informatics Campus Saarbr ucken, Germany gnad@cs.uni-saarland.de
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes The code and experimental data of our evaluation are publicly available (Gnad 2021). doi:10.5281/zenodo.4574401.
Open Datasets Yes We conducted our experiments using the Lab Python package (Seipp et al. 2017) on all benchmark domains of the International Planning Competition (IPC) from 1998-2018 in both the optimal and satisficing tracks.
Dataset Splits No The paper mentions using benchmark domains from the International Planning Competition, but it does not provide specific details on training, validation, or test dataset splits.
Hardware Specification Yes The experiments were performed on a cluster of Intel E5-2660 machines running at 2.20 GHz with the common runtime/memory limits of 30min/4Gi B.
Software Dependencies No The paper mentions using the 'decoupled search planner by Gnad & Hoffmann (2018)', 'Fast Downward planning system (Helmert 2006)', and 'Lab Python package (Seipp et al. 2017)', but does not provide specific version numbers for these software dependencies.
Experiment Setup Yes For optimal planning, we run blind search and A with h LM-cut (Helmert and Domshlak 2009); in satisficing planning, we use greedy best-first search (GBFS) with the h FF heuristic without preferred operator pruning (Hoffmann and Nebel 2001); to prove unsolvability, we run A with the hmax heuristic (Bonet and Geffner 2001).