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 Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Improving Continuous-time Conflict Based Search
Authors: Anton Andreychuk,Konstantin Yakovlev,Eli Boyarski,Roni Stern11220-11227
AAAI 2021 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate the effect of the suggested enhancements by running experiments both on general graphs and 2k-neighborhood grids. |
| Researcher Affiliation | Collaboration | 1 Peoples Friendship University of Russia (RUDN University) 2 Federal Research Center for Computer Science and Control of Russian Academy of Sciences 3 HSE University 4 Ben-Gurion University of the Negev 5 Palo Alto Research Center |
| Pseudocode | Yes | Algorithm 1: Low-level search for CCBS with DS |
| Open Source Code | Yes | Our implementation and all the raw results are available at: github.com/PathPlanning/Continuous-CBS. |
| Open Datasets | Yes | This map is publicly available in the Moving AI MAPF benchmark (Stern et al. 2019). |
| Dataset Splits | No | The paper describes generating 'scenarios' and 'MAPFR instances' for evaluation but does not specify a distinct validation set or explicit dataset split percentages (e.g., train/validation/test splits). |
| Hardware Specification | No | The paper does not provide specific details about the hardware used for the experiments, such as CPU/GPU models or memory specifications. |
| Software Dependencies | No | The paper mentions tools and algorithms like 'OMPL' and 'Dijkstra s algorithm' but does not specify version numbers for any software dependencies, which are required for reproducibility. |
| Experiment Setup | Yes | In the conducted experiments all agents were assumed to be disk-shaped with radius equal to 2/4. In each run of the evaluated algorithm, we recorded the runtime, the number of expanded CT nodes, and whether the algorithm was able to ο¬nd a solution under a time limit of 30 seconds. ... For each roadmap, 25 different scenarios were generated. Each scenario is a list of start-goal vertices, chosen randomly from the graph. Then, we pick the ο¬rst n = 2 startgoal pairs and create a MAPFR instance for n agents. If the evaluated algorithm solves this instance within the 30 seconds time limit, we proceed by increasing n by 1 and creating a new MAPFR instance. This is repeated until the evaluated algorithm is not able to solve the instance in 30 seconds. |