Hypergraph Neural Network-Based Combinatorial Optimization
基于超图神经网络的组合优化
Here you can find the code for HypOp, a tool for combinatorial optimization that employs hypergraph neural networks. It is versatile and can address a range of constrained optimization problems. In the current version, we have included the following problems: graph and hypergraph MaxCut, graph MIS, SAT, and Resource Allocation (see paper for details). To add new problems, add the appropriate loss function in the loss.py file and add the appropriate function in data_reading.py to read your specific dataset.
在这里您可以找到 HypOp 的代码，HypOp 是一种采用超图神经网络的组合优化工具。它用途广泛，可以解决一系列约束优化问题。在当前版本中，我们包含了以下问题：图和超图 MaxCut、图 MIS、SAT 和资源分配（详细信息请参阅论文）。要添加新问题，请在 loss.py 文件中添加适当的损失函数，并在 data_reading.py 中添加适当的函数以读取特定数据集。