Allocating Planning Effort When Actions Expire

Authors: Shahaf S. Shperberg, Andrew Coles, Bence Cserna, Erez Karpas, Wheeler Ruml, Solomon E. Shimony2371-2378

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

Reproducibility Variable Result LLM Response
Research Type Experimental Our empirical results indicate that our algorithms are effective in practice. This work advances our understanding of how heuristic search planners might address realistic problem settings. The empirical results suggest that taking estimated node expiration times into account can lead to a better planning strategy. When testing our algorithms, we use scenarios based on realistic search trees generated by OPTIC (Benton, Coles, and Coles 2012), the same planner that was adapted in the experiments of Cashmore et al. (2018).
Researcher Affiliation Academia 1Ben-Gurion University, Israel; 2King s College London, UK; 3University of New Hampshire, USA; 4Technion, Israel; 5UMass Lowell, USA
Pseudocode Yes Algorithm 1: Scheduling for Diminishing Returns
Open Source Code No The paper does not provide any explicit statement about releasing source code or a link to a code repository.
Open Datasets No The paper describes generating data based on distributions and search trees (from OPTIC and Robocup Logistics League), but it does not provide concrete access information (e.g., a link, DOI, or specific citation) for a publicly available dataset used in the experiments.
Dataset Splits No The paper mentions running algorithms for 500 attempts but does not specify any training, validation, or test dataset splits (e.g., percentages, sample counts, or named splits).
Hardware Specification No The paper does not provide any specific hardware details such as GPU models, CPU types, or memory specifications used for running the experiments.
Software Dependencies No The paper mentions 'OPTIC planner' and 'Bellman equation' but does not specify any software names with version numbers (e.g., Python 3.8, PyTorch 1.9).
Experiment Setup Yes We tested Schedule DR using th P t0.2, 0.5, 0.7, 1u and greedy with α P t0, 0.2, 0.5, 1, 20u, td P t1, 5, 10u. However, the reported results include only the best configuration for each algorithm: th 0.5, alpha 0, and td 1.