An adaptive optimization algorithm based on clustering analysis for return multi-flight-phase of VTVL reusable launch vehicle

Abstract

The return multi-flight-phase trajectory optimization of reusable launch vehicle is the key problem of vertical takeoff and vertical landing (VTVL) technologies. In order to solve the problem of ac-curate pinpoint soft landing in the high-dimensional parameter space composed of complex multi-variables and multi-constraints, this paper proposes a strategy of automatic parameter identification and classification and adaptive allocation of optimization algorithms based on clustering analysis. Firstly, state transition matrix of the terminal constraint variables relative to the design variables in the high-dimensional parameter space is obtained, which describes the sensitivity of the terminal constraint variables to the perturbation of the design variables. Then, K-means clustering algorithm is used to classify the sensitivity of the data elements in the state transition matrix, and the linear and nonlinear constraint variables and design variables together with linear state transition matrix are obtained according to the given linearity criterion. Finally, different optimization algorithms are allocated adaptively according to the linear and nonlinear characteristics of different parameter variables. Numerical simulation shows that the proposed algorithm can effectively deal with complex landing constraints, which has well accuracy, convergence and computational efficiency, and reduces the dimension and complexity of parameter space for the return trajectory optimization problem. Besides, the proposed algorithm is suitable for different return missions and has good adaptability and high expansibility.