Robust optimization model for water quality management in river systems

Abstract

An advanced robust optimization (RO) model to manage water quality in river systems is proposed which embraces every conceivable uncertainty present in water quality problems and is applicable to a network of streams, as an improvement over our earlier RO model. Biochemical oxygen demand (BOD) and dissolved oxygen (DO) concentrations of the incoming water from the upstream ends are added in a family of basic uncertainty parameters. In order to widen a spectrum of feasible solutions, the number of constraints associated with in-stream water quality standards is reduced. An optimization problem of BOD loading to a river system is solved with consideration to the solution robustness related to variability of the BOD load and to the model robustness related to violations of the in-stream water quality standards prescribed. Through an optimization example, it is demonstrated that the model proposed successfully provides a series of alternative strategies for water quality management in a multitudinously delimited network of streams.