Collaborative Optimization Algorithm Based on the Penalty Function

The paper discusses the collaborative optimization problems with bounded. Based the penalty function the system-level optimization convert to a unconstraint programming. To the discipline-level optimization, the normalized weighted coefficients are used and combine relaxation factors to solve. It uses the relaxation factor to expand the feasible region, and possibly makes the iteration in the calculation process run inside feasible region. The data have shown that the algorithm has expanded the choice range of the initial points with high calculation accuracy and better algorithm stability.