ABSTRACT
One of the main requirements of seismic design codes must be its easy application by structural engineers. Most of the seismic design codes are based on deterministic parameters that are focused on satisfying deterministic constraints. Nevertheless, due to the uncertain nature of the earthquakes and its effects on the structures, these design criteria cannot be considered as the best alternative to solve the problem of seismic design. A more realistic optimum design must take into account all necessary random parameters including the probability of failure or the seismic reliability of the structure. In the last decades numerous researchers have developed methods for solving structural reliability problems (Royset et al., 2001, Kumar et al., 2007). Although structural reliability methodologies are becoming widespread, serious obstacles have been encountered in practical implementations. The use of practically-applicable models or simplified models as Single-DegreeOf-Freedom (SDOF) systems representing the structural behavior of Multi-Degree-Of-Freedom (MDOF) systems associated to similar annual rate of exceeding a performance parameter is a good alternative for practical implementation in structural reliability methodologies. In the field of reliability-based seismic design of steel structures, Bojórquez et al. (2005) found probabilistic response
for the implementation of ANN to any problem is that they are not defined by means of a specific equation form; although, ANNs can be used as via to find another set of equations that can represent a specific problem. In this context, the aim of the present study is to develop a set of applicable equations for assessment of probabilistic response transformation factors using ANNs, and it is oriented at developing practical tools for application in new reliability-based seismic design criteria of steel framed buildings. In the following section some basic concepts of Artificial Neural Networks are discussed. Detailed information of ANN can be found in Haykin (1999) and Martín del Brío & Molina (2002).
