Comparison of Linear and Non-Linear Optimization Models for Storm Water and Non-Point Source Pollution Best Management Practice Decision Support

Distributed approaches to studying watershed best management practice decisions are becoming more common. Because finding global optimal solutions to non-linear systems is often time consuming, even with the use of sophisticated algorithms, the now classic techniques of linear programming may still offer some utility in the context of non-point source pollution control decision support. Two distributed linear model analogs are developed for existing non-linear distributed watershed models in order to compare best management practice (BMP) representation, optimization techniques for BMP placement decisions, computational burden, and resulting decisions. Advantages and disadvantages of these two types of models are discussed.