Overview of recent developments in hierarchical techniques for optimized control of water systems

Since water networks contain storage the optimization problem reduces to minimization of electricity charges and associated costs for the complete network over the entire control period. This can be achieved by control of pumping and storage whilst catering for consumer demands and maintaining desired reservoir levels. An optimization procedure is required which will generate 24 hour control schedules taking account of: control actions, system dynamics, physical constraints, and operating costs. Optimization techniques include dynamic programming, for single reservoir systems, and combined hierarchical methods, for multi-source multi-reservoir systems. In the paper the following non-classical hierarchical optimisation methods are outlined: primal approach with discrete and continuous controls; primal-dual approach with discrete and continuous controls; and dual approach with discrete controls and disconnected decision sets. One classical approach is considered: dual method with continuous controls