Abstract Action Scheduling for Optimal Temporal Planning via OMT
Authors: Stefan Panjkovic, Andrea Micheli
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 5 Experiments, Table 1 reports the coverage results for the considered approaches and Figure 5 (left) shows a cactus plot of the experiments, We compare against the solvers presented in (Panjkovic and Micheli 2023)..., We experimented with a version of our solver without this optimization and performed an ablation study. |
| Researcher Affiliation | Academia | 1 Fondazione Bruno Kessler, Trento, Italy 2 University of Trento, Trento, Italy {spanjkovic, amicheli}@fbk.eu |
| Pseudocode | Yes | Algorithm 1: Optimal Planning via OMT |
| Open Source Code | Yes | The solver and benchmarks are available in (Panjkovic and Micheli 2024). |
| Open Datasets | Yes | We experiment on Temporal IPC domains from the IPC14 competition, and consider two versions for each domain, one with makespan and the other with action cost minimization. We also included the benchmark set that was used in (Panjkovic and Micheli 2023). |
| Dataset Splits | No | The paper mentions using 'Temporal IPC domains from the IPC14 competition' and a 'benchmark set', but it does not specify any training, validation, or test dataset splits (e.g., percentages, sample counts, or k-fold cross-validation) for these problem instances. |
| Hardware Specification | Yes | We executed all the experiments on a cluster of identical machines equipped with Xeon E5-2440 2.4GHz and running Ubuntu Linux 20.04. |
| Software Dependencies | No | We implemented the presented approaches in C++ using the Z3 OMT solver (Bjørner, Phan, and Fleckenstein 2015). (No specific version number for Z3 is given within the paper's text). |
| Experiment Setup | Yes | We used a timeout of 3600s and a memory limit of 20GB. and The second technical optimization augments the objective criterion passed to the OMT solver with a secondary objective, consisting of minimizing the number of modf variables set to true. |