Linear stochastic optimization applied to biochemical oxygen demand – dissolved oxygen modelling

Abstract

A methodology for reflecting stochastic considerations in an optimization model is presented. The technique, which uses chance-constrained programming, is applied to a water quality management problem wherein concern is with the interaction between biochemical oxygen demand (BOD) and the dissolved oxygen (DO) concentration in a river. The uncertainty in the problem is considered to be embodied in transfer coefficients for which a lognormal distribution is derived from moment estimates provided by first-order uncertainty analysis. The appropriateness of the lognormal distribution is confirmed by results from a simulation modelling exercise. Key words: water quality, optimization, uncertainty, mathematical modelling.