INTEGRATING ROBUSTNESS INTO MULTIOBJECTIVE SPACE-VEHICLE DESIGN PROCESS

The NASA Johnson Space Center employs Monte Carlo analysis incorporate d within the Simulation and Optimization of Rocket Trajectories progra m to analyze trajectories of space vehicles. A basic function of this analysis is to assess the dispersion of a trajectory from a prescribed one due to dispersions of various vehicle and environmental parameter s. The satellite design performance is evaluated using a large number of Monte Carlo simulations. Since this method is computationally very expensive, only a few design alternatives can be evaluated. In a previ ous paper, we argued that the computational efficiency of the design p erformance evaluation could be substantially improved by replacing the Monte Carlo simulations with a simulation technique based on orthogon al arrays. In this paper we have integrated this simulation technique into a multiobjective decision support environment using a compromise decision support problem formulation. This allows us to minimize the s ystem variation while concurrently maximizing the achievement of other design goals, thus improving design quality by enabling a rational tr ade-off between nominal design performance and robustness.