Comparison of genetic programming with conventional methods for fatigue life modeling of FRP composite materials

Genetic programming is used in this paper for modeling the fatigue life of several fiber–reinforced composite material systems. It is shown that if the genetic programming tool is adequately trained, it can produce theoretical predictions that compare favorably with corresponding predictions by other, conventional methods for the interpretation of fatigue data. For the comparison of results, curves produced by the genetic programming tool are plotted together with curves produced by three other commonly used methods for the analysis of composite material fatigue data: linear regression, Whitney’s Weibull statistics and Sendeckyj’s wear-out model. The modeling accuracy of this computational technique, whose application for this purpose is novel, is very high. The proposed modeling technique presents certain advantages compared to conventional methods. The new technique is a stochastic process that leads straight to a multi-slope S–N curve that follows the trend of the experimental data, without the need for any assumptions.