Multi-Agent Path Finding Benchmark Maps
收藏资源简介:
Multi-Agent Path Finding Benchmark Maps是由卡内基梅隆大学机器人研究所创建的数据集,用于评估多智能体路径寻找算法的性能。该数据集包含33个不同大小、布局和难度的地图,旨在全面测试算法在各种场景下的表现。数据集的创建过程结合了Quality Diversity算法和神经元细胞自动机(NCA),以生成多样化的地图。这些地图广泛应用于自动化仓库、视频游戏、无人机交通管理等领域,旨在解决多智能体路径规划中的公平比较问题。
The Multi-Agent Path Finding Benchmark Maps dataset was created by the Robotics Institute of Carnegie Mellon University to evaluate the performance of multi-agent path finding algorithms. This dataset comprises 33 maps with varying sizes, layouts and difficulty levels, aiming to comprehensively test algorithm performance across diverse scenarios. The development of this dataset integrates the Quality Diversity algorithm and Neural Cellular Automata (NCA) to generate a diverse collection of maps. These maps are widely applied in domains such as automated warehouses, video games, UAV traffic management and more, with the goal of addressing the fair comparison problem in multi-agent path planning.




