Study on serial collaborative optimization algorithm based on Nelder-Mead algorithm

In aircraft multidisciplinary design optimization, Multi-method Collaborative Optimization Algorithm (MCOA) has been proposed because of the complex mathematical models. However, not only the optimization result is not very precise with general serial collaborative optimization algorithm of MCOA, but also the convergence has not been rigorously proven. In order to improve the accuracy and robustness, a method of serial collaborative optimization again algorithm based on genetic algorithm and Nelder-Mead algorithm was presented in this paper, in brief, which is first genetic algorithm then Nelder-Mead algorithm then genetic algorithm again in a circulation. Its global convergence was analyzed, accuracy and robustness was improved by a test of three typical optimization functions compared with the single algorithm and general serial collaborative optimization algorithm. The experiment results demonstrate that the method is feasible and effective.