Aircraft design using collaborative optimization
I. P. Sobieski, Ilan M. Kroo · 34th Aerospace Sciences Meeting and Exhibit · 1996
Collaborative optimization is a recently developed method for solving multidisciplinary design problems. It preserves a degree of disciplinary autonomy and allows parallel optimization. In this paper collaborative optimization is applied to two aircraft design problems to illustrate its implementation and performance characteristics. The first example, a relatively simple wing design problem, is solved using both collaborative and sequential optimization. The results are compared with those obtained using a direct, single-level optimization strategy. Collaborative optimization is shown to converge to the correct solution, while sequential disciplinary suboptimization converges to a nonoptimal result. The second example, an aircraft sizing problem, involves static aeroelastic feedback between disciplines and more complex analysis codes within the subproblems. The complete problem, in forty-five design variables and twenty-six constraints, was solved using both collaborative and single-level optimization. Both methods converge to the correct solution. The performance of the optimization strategy and the effect of increasing problem dimensionality is discussed.