The tecniques using Artificial Intelligence are more and more widespread in diferent engineering fields. Specifically, those based on Connectionist Systems and on Evolutionary Computation have been applied successfully in the analysis of big experimental data bases which were obtained in civil engineering.
In this paper it is presented an approach by means of thee two techniques applied to specific matters of reinforced concrete. The application has a bearing on the analysis of the drying creep of the concrete under controlled conditions (in specimen). The data base from RILEM is used as the main core of the test set. This article shows the process that has been done. This process is divided into three parts. First the Artificial Neural Networks (ANNs) technique has been used to identify variables that are "essential" to the phenomenon of creep. Then, with the results obtained by the ANN, an analytical study of the temporal behavior of the whole series was made. Finally, the Genetic Programming (GP) technique was applied.
The result of the process described is an equation that yields a better fit compared to the existing ones in codes and international recommendations. In addition, the influences of the dirrerent variables in the phenomenon are analyzed from the "virtual laboratory" creadted by these techniques.