Connected Static Graph Topologies

该数据集包含了5万个连接的静态图拓扑，每个图包含20个节点，用于训练多智能体强化学习（MARL）方法。该数据集不仅包括训练图拓扑，还包括测试图拓扑，旨在评估在多智能体系统中分散信息传播方法的表现。具体规模为：5万个训练图、100个包含20个节点的测试图以及100个包含50个节点的测试图。该数据集的任务是研究多智能体系统中的信息传播。

This dataset comprises 50,000 connected static graph topologies, each consisting of 20 nodes, for training multi-agent reinforcement learning (MARL) approaches. In addition to the training graph topologies, it also incorporates test graph topologies, which are designed to evaluate the performance of decentralized information propagation methods in multi-agent systems. Specifically, the dataset has the following scale: 50,000 training graphs, 100 test graphs with 20 nodes, and 100 test graphs with 50 nodes. The core task of this dataset is to study information propagation in multi-agent systems.