Some advantages of stochastic methods in multicriteria optimization of multibody systems (Q1977065)

The paper deals with multicriteria optimum design of multibody systems by using the optimization concept based on either a deterministic gradient-based method or a stochastic simulated annealing optimization procedure. The author apply the results to vehicle dynamics, and compare different optimization procedures. It turns out that simulated annealing not only avoids local minima, but also gives a clear picture of Edgeworth-Pareto optimal set and allows to estimate the influence of nonsensitive parameters. Beside the efficiency, these properties have to be considered when one decides whether a stochastic or a deterministic optimization algorithm should be chosen.