Diversity-enhanced particle swarm optimizer and its application to optimal flow control of sewer networks

This study proposes a novel diversity-enhanced Particle Swarm Optimization (PSO) scheme for optimal control of overflow in multi-reservoir sewer networks. To this aim, first two diversity boosting methodologies, cyclic neighborhood-based learning mechanism and three-phase velocity control mechanism for the particle's behavior, are proposed, and then these are combined with the constrained PSO method. Next, the linearized mathematical models of the components composing sewer networks with storage facilities are presented, and then the corresponding formulation suitable for a model predictive control (MPC) technique is discussed. Finally, the developed PSO scheme is applied to this MPC problem for minimizing the total overflow in sewer networks under a certain external inflow scenario.