Solution of Pavement Deterioration Equations by Genetic Algorithms

Abstract

Pavement deterioration models are indispensable for many purposes; as a result, a number of models are in use. Models with simple equation forms are easier to use, but frequently such models may not suffice for many purposes. Consequently, complex nonlinear forms of models are to be considered. However, determination of the solution to a complex model form is not an easy task. There are various methods of obtaining solutions to such models, with each method having its own advantages and disadvantages. The use of genetic algorithms for model development is examined in this study. A very brief description of genetic algorithms is included, and their application for the development of a model is illustrated. Five models of varied complexities, extracted from the literature, are employed to create databases in which the relationship between the response and the predictor variables is known. The solutions to the models are developed employing genetic algorithms. The results indicate a high degree of accuracy, which suggests that genetic algorithms are useful as a tool for development of solutions to pavement deterioration models.