The Road Map To Apply Evolutionary Intelligence To Asphalt Pavement Modelling

Abstract

Modelling the service life for asphalt pavements is widely made with base on empirical methods, developed half-century ago, with poor results. The new reality of the XXI century, with high construction costs, environmental restrictions and growing volumes of vehicles in the highways, enforce the shift to a new level of quality and accuracy to predict the service life of the pavements. This paper presents some insights either theoretical or experimental into a making from the ground, of an approach to predict and model the asphalt pavement behaviour using soft computing tools and, at the same time, create a way to accumulate the knowledge in this engineering field. The knowledge about the asphalt pavement life cycle is organized in a hierarchical way in order to be reused in a formal way, leading to an evolutionary process of adaptation and construction (Neves et al. 2007).

Evolutionary intelligence in asphalt pavement modeling and quality-of-information

“…Until now, the INTELLIPave method [2,31] was our proposal for asphalt pavement modeling, where a very large amount of data was used in order to build a model based on AI-based techniques. That data were collected using sensors over large periods of time (10 or more years).…”

“…The consensus tells us about an intrinsic viscous-elastic-plastic non-linear and anisotropic behavior; very complex indeed, but not sufficient to explain the performance, suggesting the existence of more undiscovered parcels of the behavior or the inexistence of any kind of pattern. Apart from the intrinsic ones, the asphalt pavement behavior is subject to the interference of many external factors, such as the characteristics of the vehicles (e.g., speed, loads) and the environmental ones (e.g., temperatures, rainfall) [1,2].…”

“…In this work we present an evolutionary system, which outcomes go far behind the ones of the INTELLIPave [31] that was our first approach to asphalt pavement modeling. It allows us to predict the asphalt pavement behavior and service-life for multi-failure criteria, in particular when applied to scenarios with incomplete information [2,31]. It is based on an extension to logic programming (ELP), and presents itself as a pure symbolic system that adapts on the fly to the changes in the environment.…”