PEA*+IDA*: An Improved Hybrid Memory-Restricted Algorithm

Authors: Frederico Messa, André Grahl Pereira10291-10298

AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Later, we perform an experimental evaluation using three memory limits and show that, compared to A +IDA on classical planning domains, PEA +IDA has higher coverage and expands fewer nodes.Empirical Analysis In this section, we aim to understand better A +IDA and PEA +IDA . Thus, we compare them using three different memory limits. We measure time as the number of expanded nodes because it avoids differences that result from implementation details.
Researcher Affiliation Academia Frederico Messa, Andr e Grahl Pereira Federal University of Rio Grande do Sul, Brazil {frederico.messa, agpereira}@inf.ufrgs.br
Pseudocode Yes Algorithm 1: PEA +IDA
Open Source Code No The paper does not contain any statement or link indicating that the source code for the described methodology (PEA+IDA) is publicly available.
Open Datasets Yes We use the STRIPS (Nilsson and Fikes 1971) optimal benchmark of 1877 tasks of the International Planning Competition (IPC).
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, or detailed splitting methodology) for training, validation, or test sets.
Hardware Specification Yes We ran all experiments with a Ryzen 3900X
Software Dependencies No The paper mentions using "the Fast Downward (Helmert 2006) framework to implement all our algorithms," but it does not specify any version numbers for this framework or other software dependencies.
Experiment Setup Yes We compare the algorithms using the remaining 332 tasks limiting the Open size to 10%, 50%, and 90% of the A s Open size peak. We use h LMCut in all the remaining experiments. all algorithms use as tie-breakers for Open first lower h-value followed by the greater depth and finally lower generation order.