Systems and methods using probabilistic forecast for agent-based control of sewers

A method for managing a sewer system, said method comprising:

• obtaining a probabilistic distribution forecast for a variable for each of multiple time intervals;
• building for said variable, based upon said probabilistic distribution forecast, multiple time interval ensemble forecasts, each ensemble forecast for one of said multiple time intervals;
• retrieving or developing a flow simulation to simulate flow of a fluid inside said sewer system;
• initializing said flow simulation, by accounting for one or more flow condition attributes present inside said sewer system, to generate an initialized flow simulation;
• running, based on an assumed value of one or more flow condition attributes present at a treatment scheme located outside said sewer system, said initialized flow simulation that uses each of said time interval ensemble forecasts to arrive at number of times one or more different potential outcomes are realized;
• establishing, based on number of times one or more of said potential outcomes are realized, a cost function of realizing one or more potential outcomes of said sewer system;
• establishing, based on one or more of said flow condition attributes in said sewer system, a cost function of flow inside said sewer system;
• and
• controlling, using a flow controller, flow of said fluid exiting said sewer system to minimize a difference between said cost function of realizing one or more potential outcomes of said sewer system and said cost function of flow inside said sewer system.