Notice: This website is continuously updated as new conference proceedings and journal volumes are published. Consequently, the figures presented here may differ from those reported in our publications.
A brief overview of our methodology is provided on the Methods page, along with references to publications documenting our approach and links to the code and data used in each study.
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.
f-Aware Conflict Prioritization & Improved Heuristics For Conflict-Based Search
Authors: Eli Boyarski, Ariel Felner, Pierre Le Bodic, Daniel D. Harabor, Peter J. Stuckey, Sven Koenig12241-12248
AAAI 2021 | Venue PDF | LLM Run Details | Input Tokens: 17,203 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 4,841 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental Results All our experiments were run on a Linux machine with an Intel Xeon CPU E5-2630 v2 running at 2.60GHz, with memory usage limited to 8GB. We used Gurobi 8.1.1 to solve the MIP models of the MVCs of the various heuristics. We experimented on the MAPF benchmark (Stern et al. 2019), which contains 32 grids of different types (city maps, grids with random obstacles, mazes, warehouse maps, etc.), each with 25 scenarios that specify the start and target locations for up to 7,000 agents. |
| Researcher Affiliation | Academia | 1Ben-Gurion University of the Negev 2Monash University 3University of Southern California |
| Pseudocode | No | The paper describes methods and processes but does not contain any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Our code is available at https://github.com/eli-b/fcardinal. |
| Open Datasets | Yes | We experimented on the MAPF benchmark (Stern et al. 2019) |
| Dataset Splits | No | The paper refers to the 'MAPF benchmark (Stern et al. 2019)' and its 'scenarios' but does not specify explicit train/validation/test dataset splits (percentages, counts, or specific methods like k-fold cross-validation) for their experiments. |
| Hardware Specification | Yes | All our experiments were run on a Linux machine with an Intel Xeon CPU E5-2630 v2 running at 2.60GHz, with memory usage limited to 8GB. |
| Software Dependencies | Yes | We used Gurobi 8.1.1 to solve the MIP models of the MVCs of the various heuristics. |
| Experiment Setup | Yes | We increased the number of agents on these scenarios until we reached the runtime limit of 60 seconds. For higher numbers of agents, the solver was considered to have failed implicitly. The baseline solver in all of our experiments was only aware of g-cardinal conflicts, used the CG heuristic, and implemented bypassing conflicts (Boyarski et al. 2015a). |