Optimization algorithm of the surrogate model structure based on pigeon swarm search strategy

The hypersonic vehicle has problems with multiple systems coupling and multivariate design. Hence, it is difficult to determine the surrogate model structure of the hypersonic vehicle under larger samples. To solve this problem, this paper proposes a surrogate model structure optimization method based on the idea of PIO algorithm. The method can independently search for a polynomial surrogate model structure that satisfies the accuracy and prediction performance requirements. Through comparative analysis, utilizing this method outperforms Davidson’s genetic programming algorithm. Furthermore, it was applied to the global aerodynamic data fitting and control design of a hypersonic vehicle, and the simulation results reveal that the error of the prediction data obtained by this method was less than 5%, which has a good fitting effect and can meet the engineering application requirements.