An adaptive approach to constraint aggregation using adjoint sensitivity analysis

摘要: Constraint aggregation is the key for efficient structural optimization when using a gradient-based optimizer and an adjoint method sensitivity analysis. We explore different methods of constraint numerical optimization. analyze existing approaches, such as considering all constraints individually, taking maximum Kreisselmeier–Steinhauser (KS) function. A new adaptive approach based on KS function proposed that updates parameter by into account sensitivity. This shown to significantly increase accuracy results without additional computational cost especially large number are active at optimum. The characteristics each performance solving wing structure weight minimization problem.