Energy cost optimization in water distribution systems using Markov Decision Processes

A major concern for much of the world today is the scarcity of potable water and energy. The rising demand for water and the need for more electricity to transport it, from remote sources to the consumers, are pushing humanity to find ways to save water as well as energy. A way to deal with those two problems and that also possess an encouraging financial return is to have more efficient operations in water distribution systems. In this paper we propose a novel framework to optimize the electricity cost of water utility companies using Markov Decision Processes. We report experiments in a small-sized city demonstrating that our methodology is effective for reducing the electricity expenses and for keeping the water pressure in a controlled range. This new technique will prevent water outages and also minimizes the risk of pipe rupture, besides increasing the reliability of the entire distribution system by reducing inefficiencies in the operation.