A genetic algorithm for the shape optimization of parts subjected to thermal loading

In many technical situations, the optimization of the mechanical behavior o f structures proceeds from the search for the ideal shape satisfying therma l, mechanical, technological and geometrical constraints. In this article, the shape optimization of mono- and two-dimensional structures is handled b y means of a new genetic algorithm (GA). The method is in general well suit ed to the resolution of nonconstrained optimization problems: The algorithm presented here has been modified by taking into account the imposed design constraints in the selection of the "individuals" belonging to a gives pop ulation. The crossover operation between individuals and the mutation proce ss in their original forms are applied to derive the optimal shape of parts subjected to thermal loadings. The algorithm exhibits a good convergence t oward the optimal solation and the numerical results of its application sho w a good numerical accuracy.