Automatic Generation of Flexible Plans via Diverse Temporal Planning

Authors: Yotam Amitai, Ayal Taitler, Erez Karpas6049-6057

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

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, an empirical evaluation on a set of IPC benchmarks shows that our approach scales well, and generates TPNs which can generalize the set of plans they are generated from.
Researcher Affiliation Academia Yotam Amitai, Ayal Taitler, Erez Karpas Technion Israel Institute of Technology yotama@campus.technion.ac.il, {ataitler, karpase}@technion.ac.il
Pseudocode No The paper describes procedures narratively and with figures, but does not include any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not provide a statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets Yes We evaluated our approach on domains from the temporal track IPC in 2011, 2014, and 2018.
Dataset Splits No The paper uses IPC benchmarks but does not specify how the dataset was split into training, validation, or test sets, nor does it provide percentages or sample counts for these splits.
Hardware Specification Yes The experiments were performed on Intel i7-7700K 32GB RAM
Software Dependencies Yes We then construct a COP based on M in Minizinc (Nethercote et al. 2007) and use Gecode (Schulte, Tack, and Lagerkvist 2019) to solve it. ... The Fast Downward planning system (Helmert 2006) extended with the support for structural symmetries and the orbit space search algorithm (Domshlak, Katz, and Shleyfman 2015) with LM-cut heuristic (Helmert and Domshlak 2009) as was used in (Katz et al. 2018).
Experiment Setup Yes We evaluated all combinations of k chosen from {2, 4, 8}, both compatibility methods (Full & Semi denoted as F,S) and merging transitivity (Strict & Loose denoted as St,L). Thus, for each planning problem, we ran the diverse planner 3 times (for the different values of k), and then ran 4 different versions of our COP (for the different compatibility and transitivity). The experiments were performed on Intel i7-7700K 32GB RAM, with a time limit of 30min for generating the diverse solutions and 30min for the COP task.