Can I Really Do That? Verification of Meta-Operators via Stackelberg Planning
Authors: Florian Pham, Alvaro Torralba
IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate meta-operators on a standard set of IPC benchmarks. All experiments were conducted on a cluster of Intel Xeon CPU E5-2660 with 2.20GHz using Downward Lab [Seipp et al., 2017]. Each run had a time limit of 1800 seconds and a memory limit of 4 GB. |
| Researcher Affiliation | Academia | 1Saarland University, Saarbr ucken, Germany 2Aalborg University, Aalborg, Denmark |
| Pseudocode | Yes | Table 1: Action definition for all actions in the Stackelberg planning task to verify if d is a valid meta-action. In the effect definition, positive literals means that they are add effects, whereas negative literals represent delete effects. As syntactic sugar, we use conditional effects of the form (cond eff ), where the effect eff is applied if and only if the cond holds in the state where the action is applied. |
| Open Source Code | Yes | The code and data are publicly available [Pham and Torralba, 2023]. |
| Open Datasets | Yes | We evaluate meta-operators on a standard set of IPC benchmarks. We use a small set of 5 instances per domain generated with the publicly available random generators of these IPC domains. |
| Dataset Splits | No | The paper mentions different sets of instances for training, validation, and evaluation (e.g., 'training instances', 'validation instances', 'IPC instances not used for learning or validating'), but does not explicitly provide specific percentages, sample counts, or detailed methodologies for these splits in the main text. |
| Hardware Specification | Yes | All experiments were conducted on a cluster of Intel Xeon CPU E5-2660 with 2.20GHz |
| Software Dependencies | No | The paper mentions 'Downward Lab [Seipp et al., 2017]' and 'Fast Downward planning system [Helmert, 2006]' but does not provide specific version numbers for these or other relevant software dependencies. |
| Experiment Setup | Yes | We use two configurations that differ on whether the plans computed for the follower are optimal or not. The SLS-opt configuration, used by Torralba et al. [2021], is an optimal symbolic search configuration [Torralba et al., 2017]. The SLS-sat configuration, on the other hand, uses Greedy Best-First Search with the FF heuristic [Hoffmann and Nebel, 2001]. In the SLS-sat configuration, we also modify the symbolic leader search algorithm by initializing the initial follower cost to a large constant (10^5). |