Multi-Agent Path Finding with Delay Probabilities

Authors: Hang Ma, T. K. Satish Kumar, Sven Koenig

AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments We evaluate AME with MCPs on a 2.50 GHz Intel Core i5-2450M PC with 6 GB RAM. ... Table 1 reports for each MAPF-DP instance the runtime, the approximate average makespan calculated by AME, the average makespan over 1,000 plan-execution runs with MCPs together with 95%-confidence intervals and the number of sent messages.
Researcher Affiliation Academia Hang Ma Department of Computer Science University of Southern California hangma@usc.edu; T. K. Satish Kumar Department of Computer Science University of Southern California tkskwork@gmail.com; Sven Koenig Department of Computer Science University of Southern California skoenig@usc.edu
Pseudocode Yes Algorithm 1: High-Level Search of AME.
Open Source Code No The paper does not provide any concrete access information (e.g., repository link, explicit statement of code release) for the methodology described.
Open Datasets No The paper describes generating its own MAPF-DP instances (e.g., '10 MAPFDP instances (labeled random 1-10) in 30 30 4-neighbor grids', '10 MAPF-DP instances (labeled warehouse 1-10) in a simulated warehouse environment') but does not provide concrete access information (link, DOI, formal citation) to these generated datasets being made publicly available.
Dataset Splits No The paper describes generating MAPF-DP instances for evaluation but does not specify any training, validation, or test dataset splits. The term 'validation' in the paper refers to the 'validation' of MAPF-DP plans and not dataset splitting.
Hardware Specification Yes We evaluate AME with MCPs on a 2.50 GHz Intel Core i5-2450M PC with 6 GB RAM.
Software Dependencies No The paper mentions various frameworks and algorithms like 'CBS', 'Push and Swap', and 'POMDPs' but does not specify version numbers for any software components or libraries used for implementation.
Experiment Setup Yes We generate 10 MAPFDP instances (labeled random 1-10) in 30 30 4-neighbor grids with 10% randomly blocked cells and random but unique start and unique goal cells for 35 agents whose delay probabilities for AME are sampled uniformly at random from the delay probability range (0, 1/2).