Jeffrey Rieker (Special Studies Manager, BOR Lahontan Basin Area Office)

Abstract: This presentation will discuss a generalized reservoir release optimization mechanism based on reinforcement learning methodologies from the field of artificial intelligence, with a focus on the application to water quality issues on the Truckee River in California and Nevada. The reinforcement learning mechanism is used in conjunction with the Truckee River long-term RiverWare planning model to evaluate the optimal use of stored water for improving water quality downstream of Truckee River reservoirs. The presentation will highlight the artificial intelligence mechanism and its interaction with RiverWare through rule-fired DMI’s. Preliminary results from the Truckee River case study will also be discussed.