COMPARISON OF PROBABILISTIC AND DETERMINISTIC OPTIMIZATIONS USING GENETIC ALGORITHMS

This paper describes an application of genetic algorithms to determini stic and probabilistic (reliability-based) optimization of damping aug mentation for a truss structure. The probabilistic formulation minimiz es the probability of exceeding upper limits on the magnitude of the d ynamic response of the structure due to uncertainties in the propertie s of the damping devices. The corresponding deterministic formulation maximizes a safety margin with respect to these limits. Because this w ork was done in the context of an experimental comparison of the relia bilities of the resulting designs, antioptimization was used to maximi ze the contrast between the probabilities of failure of the two design s. This contrast maximization was also performed with a genetic algori thm. The paper describes the genetic algorithm used for the optimizati on and antioptimization, and a number of modifications to the antiopti mization formulation intended to reduce the computational expense to a n acceptable level. Optimal designs are obtained for both formulations . The probabilistic design shows a very significant improvement in rel iability.