Global and Evolutionary Optimization for Water Management Problems

Abstract

Optimization is a type of modelling which provides solutions to problems that concern themselves with the choice of a “best” configuration or a set of parameters that achieve some objective. The search for the “best” (maximum or minimum) is called global optimization where the goal is to find the best solution in nonlinear models that may have a multitude of local optima. Evolutionary optimization is a term denoting algorithms that mimic natural evolution and “evolve” better solutions through evolutionary mechanisms such as selection, reproduction, crossover and mutation. They are particularly suited for water system applications where the optimization problem often involves a mixture of discrete and continuous decision variables or potential solutions that must be evaluated with complex simulation models, whose linearization or derivative calculations would be difficult or impossible. Evolutionary optimization is a mature technology with a large number of applications in a multitude of areas, including water engineering and management. This chapter explains the basics of evolutionary optimization and illustrates its use in many areas of water system management.