Collaborative optimization algorithm and its applications to complex system parallel design

For complex system design in aeronautic and astronautic field, the conventional design strategy is computationally expensive and time-consuming. Collaborative Optimization (CO) is a promising method to solve this problem due to its parallel design strategy. CO divides the optimization problem into two levels one system-level optimization and N parallel disciplinary-level optimizations. However, the consistence constraints make the system-level optimization unsolvable. In this paper we proposed a new concept of inter-disciplinary inconsistency information based on which the consistence constraints were slacked and a hyper-ball system-level design space was constructed. Comparing with the existing relaxation of system-level constrains, tolerance determined by this method is more reasonable for it is guaranteed for the solvability and takes the consistence into account simultaneously. Moreover comparing with the exiting response surface method, this method can reduce the computational job drastically. In the last section a gear reducer box design problem was divided into 2 parts: shaft design and gear design which were parallelly performed by the proposed method. The result shows its efficiency and precision.