Plan-Length Bounds: Beyond 1-Way Dependency
Authors: Mohammad Abdulaziz7502-7510
AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We experimentally show that this additional decomposition improves the bounds computed by the algorithms from AGN1 and AGN2. For instance, bounds computed using the method from AGN2 improved in 68% of the planning problems on which we experimented, with an improvement of at least 50% in 71% of the cases. We experimentally evaluate different bounding algorithms on problems from previous International Planning Competitions (IPC), and the unsolvablity IPC, open Qualitative Preference Rovers benchmarks from IPC2006 (to which we refer as NEWOPEN) and the hotel-key protocol verification problem from AGN2. Our experiments were conducted on a uniform cluster with time and memory limits of 30minutes and 8GB, respectively. |
| Researcher Affiliation | Academia | Mohammad Abdulaziz Technical University of Munich, Munich, Germany |
| Pseudocode | Yes | Algorithm 1: TRAVD(δ) |
| Open Source Code | No | The paper does not explicitly state that source code for its own methodology is made available, nor does it provide a link to a repository. |
| Open Datasets | Yes | We experimentally evaluate different bounding algorithms on problems from previous International Planning Competitions (IPC), and the unsolvablity IPC, open Qualitative Preference Rovers benchmarks from IPC2006 (to which we refer as NEWOPEN) and the hotel-key protocol verification problem from AGN2. |
| Dataset Splits | No | The paper evaluates algorithms on 'planning problems' and 'benchmarks' from competitions but does not specify training, validation, or test dataset splits in terms of percentages, counts, or explicit partitioning methodologies. |
| Hardware Specification | No | Our experiments were conducted on a uniform cluster with time and memory limits of 30minutes and 8GB, respectively. |
| Software Dependencies | No | The paper mentions software tools like 'Madagascar MP (Rintanen 2012)' and 'Fast Downward s preprocessing step (Helmert 2006)' but does not provide specific version numbers for these or any other software dependencies. |
| Experiment Setup | No | The paper mentions experimental limits like '30minutes and 8GB' but does not provide specific details on hyperparameters, model initialization, or training settings for a typical experimental setup. |