Classic control theory offers different control strategies that are traditionally applied to several fields, such as industrial processes and robotics. The water resources management field adopted some of these techniques to design the operation of regulated lakes, a control problem that presents some peculiarities, e.g. the presence of multiple objectives, non-linearities and strong uncertain disturbances (e.g., inflows to the lake). However, when dealing with the planning of new water reservoirs, such a control problem – designing the optimal operation - must be solved jointly with a planning problem - designing the optimal water reservoir size - given their strict interdependency. These peculiar characteristics pose new challenges to the application of classic control approaches to the water resources management field. The existing literature already proposed different approaches to tackle some of these challenges (e.g., system non-linearities and uncertain disturbances), by identifying methods that require strong approximations (e.g., linear system, white disturbances) as well as more sophisticated formulations. However, approaches to handle more advanced issues have not been discussed yet. We thus propose a classification of methodological approaches to address peculiar problems of water resources systems, namely planning and management of reservoirs, use of exogenous information to capture uncertainty, and the presence of multiple objectives.

The seminar is based on the 7^th oCPS PhD School on Cyber-Physical Systems held at IMT School of Advanced Studies in Lucca.