Efficient Temporal Planning Using Metastates

Authors: Amanda Coles, Andrew Coles, J. Christopher Beck7554-7561

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our evaluation shows that this approach, implemented in OPTIC and compared to existing state-of-the-art memoisation techniques, improves performance across a range of temporal domains. We empirically evaluate our approach on temporal planning domains and our results show a significant improvement in performance over the state-of-the-art in memoisation for temporal planning.
Researcher Affiliation Academia Amanda Coles, Andrew Coles, J. Christopher Beck *Department of Informatics, King s College London, UK. Department of Mechanical & Industrial Engineering, University of Toronto, Canada. email: {amanda,andrew}.coles@kcl.ac.uk, jcb@mie.utoronto.ca
Pseudocode Yes Algorithm 1: Memoise, Algorithm 2: Add Member To Meta State, Algorithm 3: Search, Algorithm 4: Find Another Member
Open Source Code No The paper does not provide any explicit statements about releasing the source code or links to a code repository for the methodology described.
Open Datasets Yes For evaluation domains, we take International Planning Competition benchmarks and temporally interesting domains from the literature. Cafe and Driverlog Shift (Coles et al. 2009); P2P (Huang et al. 2009); TMS and Turn and Open (IPC2011); the compiled timewindows/deadlines variants of Pipes No-Tankage, Satellite and UMTS (IPC2004) and 5 other IPC domains Rovers (2002); Pipes Tankage (2004); Transport, Elevators and Openstacks (2008).
Dataset Splits No The paper mentions using International Planning Competition benchmarks but does not specify explicit training, validation, or test dataset splits (e.g., percentages or sample counts).
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions that the approach is implemented in 'OPTIC' and uses 'PDDL2.1', but it does not specify version numbers for these or any other software dependencies.
Experiment Setup Yes In both cases we use WA* search (with W=5); and a temporal RPG heuristic, as in OPTIC.