Institute: District of Columbia
Year Established: 2020 Start Date: 2020-03-01 End Date: 2021-02-28
Total Federal Funds: $9,962 Total Non-Federal Funds: $38,280
Principal Investigators: Arash Massoudieh
Abstract: This research explores the potential for using computational intelligence generally, and reinforcement learning in particular, to maximize long term benefits of urban green infrastructure through smart control. Reinforcement learning is a machine learning approach to decision making that can help make real-time decisions to maximize long term reward based on continually updated information about the state of the world and has the potential to reveal river management strategies that have not been previously discovered. This research will develop a reinforcement learning algorithm specially for optimal operation of stormwater green infrastructure where there is a clear trade-off between different competing uses, where the hydrologic future is uncertain, and where improved receiving water ecosystem is a priority. The algorithm will be applied to the Sligo Creek watershed in the vicinity of Washington, DC to discover optimal operation strategy for operating several existing stormwater pondsand several other hypothetical ones to be constructed. Broader impacts of this work include a Signiant step forward in applying artificial intelligence to critical environmental sustainability challenges generally and improving stream ecosystem in lieu of long-term change specifically.