Automated co-design of high-performance thermodynamic cycles via graph-based hierarchical reinforcement learning

This dataset is part of the study "Automated co-design of high-performance thermodynamic cycles via graph-based hierarchical reinforcement learning"(arXiv:2604.13133 [cs.LG] https://doi.org/10.48550/arXiv.2604.13133).. The file “Comparison of valid cycle generation by the HRL agent and random search.xlsx” contains the raw data corresponding to Figure 4.The file “Cycle index versus optimal performance.xlsx” contains the raw data corresponding to Figure 6.