Use Koopman emulator to train RLs. Source code and data. The published article refer to https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2021WR030939
Chaohu: A real-case model of a combined sewer system with 3 pump stations and forebays. Papers use this model:
-
Liao, Z., Gu, X., Xie, J., Wang, X., & Chen, J. (2019). An integrated assessment of drainage system reconstruction based on a drainage networkmodel. Environmental Science and Pollution Research, 26(26), 26563–26576. https://doi.org/10.1007/s11356-019-05280-
-
Zhi, G., Liao, Z., Tian, W., Wang, X., & Chen, J. (2019). A 3D dynamic visualization method coupled with an urban drainage model. Journal ofHydrology, 577, 123988. https://doi.org/10.1016/j.jhydrol.2019.123988
-
Tian, W., Liao, Z., Zhi, G., Zhang, Z.&Wang, X., 2022b. Combined Sewer Overflow and Flooding Mitigation Through a Reliable Real-Time Control Based on Multi-Reinforcement Learning and Model Predictive Control. Water Resources Research, 58(7): e2021WR030703. https://doi.org/10.1029/2021WR030703
-
Tian, W., Liao, Z., Zhang, Z., Wu, H.&Xin, K., 2022a. Flooding and Overflow Mitigation Using Deep Reinforcement Learning Based on Koopman Operator of Urban Drainage Systems. Water Resources Research, 58(7): e2021WR030939. https://doi.org/10.1029/2021WR030939
-
Zhang, Z., Tian, W., & Liao, Z. (2023). Towards coordinated and robust real-time control: A decentralized approach for combined sewer overflow and urban flooding reduction based on multi-agent reinforcement learning. Water Research, 229, 119498. https://doi.org/10.1016/j.watres.2022.119498
- tensorflow >= 2.3
- pyswmm >= 0.6.2