Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Solving Delete Free Planning with Relaxed Decision Diagram Based Heuristics
Authors: Margarita Paz Castro, Chiara Piacentini, Andre Augusto Cire, J. Christopher Beck
JAIR 2020 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our empirical analysis shows that while DD-based heuristics trail the state of the art, even small relaxed DDs are competitive with the linear programming heuristic for the DFP task, thus, revealing novel ways of designing admissible heuristics. |
| Researcher Affiliation | Academia | Margarita P. Castro EMAIL Department of Mechanical and Industrial Engineering University of Toronto Toronto, Ontario M5S 3G8, Canada Chiara Piacentini EMAIL Department of Mechanical and Industrial Engineering University of Toronto Toronto, Ontario M5S 3G8, Canada Andre A. Cire EMAIL Department of Management, University of Toronto Scarborough Rotman School of Management Toronto, Ontario M1E 1A4, Canada J. Christopher Beck EMAIL Department of Mechanical and Industrial Engineering University of Toronto Toronto, Ontario M5S 3G8, Canada |
| Pseudocode | Yes | Algorithm 1 Relaxed MDD (BDD) Construction Procedure 1: procedure Construct MDD(Π+, W, m) 2: M := Width One MDD(m) 3: while M has been modified do 4: Update MDDTop(M, W, Π+) 5: Update MDDBottom(M, W, Π+) 6: return M |
| Open Source Code | Yes | Our code is available at: https://github.com/Margarita Castro/dd-planning. |
| Open Datasets | Yes | We test all approaches over delete-free version of domains from the IPC2011 and IPC2014 competitions. |
| Dataset Splits | No | The paper uses delete-free versions of domains from IPC2011 and IPC2014 competitions. It also describes generating "10 random instances for each grid configuration" for a modified visit-all domain. However, it does not specify explicit training, validation, or test splits for any dataset used in the context of model training or evaluation splits in the typical machine learning sense. |
| Hardware Specification | No | The paper does not provide specific hardware details such as GPU/CPU models, processor types, or memory amounts used for running experiments. It mentions that "Everything is coded in C++." and "We use Gurobi 8.0 to solve the LP and MILP models.", which are software details. |
| Software Dependencies | Yes | We use Gurobi 8.0 to solve the LP and MILP models. Everything is coded in C++. |
| Experiment Setup | Yes | Our experiments consider a 30 minute time limit and a 2GB memory limit. We analyze our heuristics performance using different maximum widths W {2, 4, 8, 16, 32, 64} and compare their performance against a MILP model for DFP (Imai & Fukunaga, 2014, 2015) and its LP relaxation. |