Watershare reposted this
❓ How should the world secure high-quality water services while facing severe climate, economic and population pressures under deep uncertainty? Water distribution network (WDN) planning for the long term is complex due to the deep uncertainty that characterizes essential design parameters, such as urban development, demographic shifts, resource availability and climate. Lydia T., Christos Makropoulos and Dragan Savic FREng suggest a reinforcement learning (RL) approach to address the deterministic design of WDNs. One distinctive RL characteristic is that agents learn to adapt to unforeseen circumstances, making it a promising approach to decision-making in dynamic and uncertain environments. Through the Hanoi network benchmark, they demonstrated the feasibility of using RL for WDN design. The results showed that the agent could find cost-effective network designs that meet the minimum pressure requirements, making RL a promising approach worthy of further exploration. 'In future work, we will incorporate more complex reward and state representations to enhance the algorithm's performance' says Lydia T. https://lnkd.in/et9vvr4N #reinforcementlearning, #waterdistribution, #networks, #proximalpolicy #optimization
