多智能体超图交通建模仿真数据

该数据集为仿真数据，主要面向复杂动态系统中的多智能体超图建模与感知。原始数据基于仿真方式生成，涵盖多种典型交通场景与环境条件，每组数据由多个路口节点与道路超边组合，用于评估算法在不同交通流、时段以及智能体交互关系下的性能表现。在本研究中，为了支持复杂动态系统中的多智能体超图建模与分析，我们基于Matlab环境下运行相关仿真与数据生成脚本，对大连理工大学校园交通网络进行建模并筛选与预处理，包括对仿真输出数据进行拓扑匹配、属性归一化等操作，以保证不同要素之间的对齐精度。最终从仿真结果中选取了 40 个路口节点与 53 条道路段，并模拟各时段（早高峰、平峰、晚高峰、夜间等）下的交通流量、平均车速及智能体交互强度信息，全部保存为标准化mat格式，数据容量约为 718 KB。基于该数据集，可在多智能体协同感知、交通流预测与监测、智能调度与超图优化等多领域开展进一步实验，包括针对节点超度、边强度等加权指标的复杂交通系统研究，以及对路口协同调度策略的验证和评估。该建模仿真数据集为后续的交通系统建模与多智能体应用提供了丰富且多样的测试环境与数据支撑。This is a simulated dataset targeting multi-agent hypergraph modeling and perception in complex dynamic systems. The raw data is generated through simulation, covering multiple typical traffic scenarios and environmental conditions. Each data sample comprises multiple intersection nodes and road hyperedges, and is used to evaluate algorithm performance under different traffic flows, time periods, and agent interaction relationships. In this study, to support multi-agent hypergraph modeling and analysis in complex dynamic systems, we ran relevant simulation and data generation scripts in the Matlab environment, modeled the Dalian University of Technology campus traffic network, followed by screening and preprocessing operations including topology matching and attribute normalization on the simulation output data to ensure the alignment precision between different elements. Finally, 40 intersection nodes and 53 road segments were selected from the simulation results, and traffic flow, average speed, and agent interaction intensity information under various time periods (morning peak, off-peak, evening peak, night, etc.) were simulated. All data was saved in standard .mat format, with a total data capacity of approximately 718 KB. Based on this dataset, further experimental research can be conducted across multiple domains including multi-agent collaborative perception, traffic flow prediction and monitoring, intelligent scheduling, and hypergraph optimization, such as investigations into complex traffic systems based on weighted indicators like node hyperdegree and edge strength, as well as the validation and assessment of intersection cooperative scheduling strategies. This modeling and simulation dataset provides rich and diverse test environments and data support for subsequent traffic system modeling and multi-agent application developments.

搜集汇总

背景与挑战

背景概述

该数据集为基于Matlab仿真的多智能体超图交通建模数据，模拟大连理工大学校园交通网络，包含40个路口节点和53条道路段，覆盖不同时段的交通流量、车速及智能体交互信息。它适用于交通流预测、多智能体协同感知和超图优化等研究领域，提供标准化mat格式数据支持实验验证。