Since its inception, the long range guided rocket (LGR) has demonstrated powerful capabilities in various combat scenarios and has become one of the weapon systems that countries are competing to develop. Compared to tactical missiles, long-range guided rockets have lower costs and have a range, lethality, and strike accuracy far beyond conventional rockets. They have been widely integrated into modern weapon and equipment systems. With the improvement of weapon modernization level, the demand for high-performance, low-cost, and highly reliable design of long-range guided rockets is becoming increasingly urgent, which puts forward higher requirements for existing design methods. Remote guided rockets are complex systems composed of multiple subsystems, and their design process is characterized by multidisciplinary, strong coupling, nonlinearity, and high time consumption. However, the existing optimization design methods combined with complex multidisciplinary simulation models have low efficiency, which greatly limits the engineering applicability of remote guidance rocket optimization design. Therefore, reducing computational complexity and improving overall design optimization efficiency through approximate modeling techniques are of great significance for enhancing the performance of long-range guided rockets.

The sequence approximation optimization method was widely used in the optimization design of high time-consuming simulation models. The following proposed a sequence approximation optimization method based on improved augmented radial basis, which included experimental design, approximation modeling, and sequence sampling. The experimental design method was applied to generate a small number of samples to establish an initial approximate model with relatively low accuracy. The training set was gradually expanded by adding new samples, and the approximate model was dynamically updated. Adding sample points can not only verify the accuracy of existing approximate models, but also gradually improved the model accuracy, thereby achieving efficient prediction of the global optimal solution.

The sequence approximation optimization results converge after approximately 620 iterations. Through optimization, the range increased from the initial 940 km to 1097 km, an increase of about 16.7%, and the total engine weight decreased from 4400 kg in the initial plan to 4373 kg. The length of the optimized design cone has significantly increased, and the installation position of the tail wing has moved forward from the half span position, resulting in improved aerodynamic performance. In the parameters of the power system, the increase in the width of the propellant column wing is used to obtain a larger burning surface and improve engine thrust, while other parameters do not change significantly; In the flight trajectory parameters, the launch speed inclination angle slightly increases, and there are significant changes in the angle of attack sequence. The maximum flight altitude of the guided rocket has been increased to around 100km, with a slight increase in engine thrust, a slight increase in landing angle and terminal Mach number, a significant increase in maximum normal overload, and a small change in dynamic pressure. The changes in flight performance are consistent with the results of parameter analysis.

Established an overall performance analysis model for remote guided rockets. Construct an overall performance calculation model for remote guided rockets based on disciplines such as geometry, aerodynamics, mass, dynamics, and ballistics, provide an initial plan, and complete trajectory simulation. A sequence approximation optimization method based on improved augmented radial basis function was proposed. The efficiency of approximation modeling was improved through recursive evolution experimental design and fast cross validation, and adaptive sampling was performed based on non-precise strategies. The numerical results showed that the proposed method had significant advantages. For the multidisciplinary optimization problem of remote guided rockets, the proposed method was adopted to obtain an optimized design scheme that satisfies constraints and has superior performance. The range increased from the initial design of 940 km to 1097 km, an increase of about 16.7%.