Applying Genetic Programming to Civil Engineering in the Improvement of Models, Codes and Norms

This paper presents the use of Evolutionary Computation (EC) techniques, and more specifically the Genetic Programming (GP) technique, to Civil Engineering. This technique is applied here to a phenomenon that models the performance of structural concrete under controlled conditions throughout time. Several modifications were applied to the classic GP algorithm given the temporal nature of the case to be studied; one of these modifications was the incorporation of a new operator for providing the temporal ability for this specific case. The fitness function has been also modified by adding a security coefficient that adjusts the GP-obtained expressions and penalises the expressions that return unstable values.