Evolutionary Algorithm Memory Enhancement for Real-Time CSO Control

Control of combined sewer overflows (CSOs) may be enhanced through real-time decision support. An optimization algorithm adapted for changing rainfall is used to dynamically control complex sewer hydraulics to minimize CSO volume. Different methods of enhancing the optimization for real-time processing consist of: separating the hydraulic model for multi-objective optimization or to optimize only critical portions of the sewer system, using an efficient optimization technique, and incorporating memory into the optimization to speed convergence to a solution for each forecasted rainfall change. Potential optimal management solutions and the associated environmental characteristics can be stored and used to re-initialize the optimization at each environmental change. The memory may also be altered to indicate what precision of hydraulic model should be used for different rainfall conditions.