This work gives two examples of application of stochastic techniques for the optimization of stiffened plates or shells. The research strategy consists in substituting, for finite‐element calculations in the optimization process, an approximate response of a neural network, or an approximate response from the Ritz method. More precisely, the paper describes the use of a backpropagation neural network or the Ritz method in creating function approximations for use in computationally intensive design optimization based on genetic algorithms. Two examples of applications are presented; the first one deals with the optimization of stiffeners on a plate by varying their positions, while having well‐defined dimensions; the second example deals with the optimization of a thin shell subject to buckling. Copyright © 2001 John Wiley & Sons, Ltd.