Tree-based reinforcement learning for optimal water reservoir operation

Although being one of the most popular and extensively studied approaches to design water reservoir operations, Stochastic Dynamic Programming is plagued by a dual curse that makes it unsuitable to cope with large water systems: the computational requirement grows exponentially with the number of state variables considered (curse of dimensionality) and an explicit model must be available to describe every system transition and the associated rewards/costs (curse of modeling). A variety of simplifications and approximations have been devised in the past, which, in many cases, make the resulting operating policies inefficient and of scarce relevance in practical contexts. In this paper, a reinforcement-learning approach, called fitted Q-iteration, is presented: it combines the principle of continuous approximation of the value functions with a process of learning off-line from experience to design daily, cyclostationary operating policies. The continuous approximation, performed via tree-based regression, makes it possible to mitigate the curse of dimensionality by adopting a very coarse discretization grid with respect to the dense grid required to design an equally performing policy via Stochastic Dynamic Programming. The learning experience, in the form of a data set generated combining historical observations and model simulations, allows us to overcome the curse of modeling. Lake Como water system (Italy) is used as study site to infer general guidelines on the appropriate setting for the algorithm parameters and to demonstrate the advantages of the approach in terms of accuracy and computational effectiveness compared to traditional Stochastic Dynamic Programming.