Evolutionary programming in dowelled-concrete pavement: Faulting performance modeling

Accurate modeling of the faulting performance of concrete pavement is a critical component of pavement-management systems. A characteristic feature of most faulting models is that they are formulated and estimated statistically from field data. Most of these statistical models are obtained by trial-and-error and various transformations, and an a priori functional form is selected and applied only under static conditions. However, the availability of the large databases included in these models introduces the novel use of evolutionary programming to model the faulting performance of dowelled-concrete pavements. Evolutionary programming belongs to a new class of computational techniques, known as evolutionary computation. In this paper, an evolutionary algorithm is used to obtain a model that more closely matches faulting performance.