论文部分内容阅读
针对卫星结构要求总质量轻小、一阶固有频率较高的特点,建立了卫星结构多目标优化的数学模型,通过BP神经网络和遗传算法相结合进行参数优化,并编制了相应的计算程序,既利用了神经网络的非线性映射、网络推理和预测功能,又发挥了遗传算法的全局优化特性,得出了合理的优化结果,与传统的结构优化方法相比,此方法效率较高,精度良好。
According to the characteristics of small total mass and low natural frequency of satellite structure, a mathematical model of multi-objective optimization of satellite structure is established. The parameters are optimized by BP neural network and genetic algorithm, and the corresponding calculation program is compiled. It not only takes advantage of the nonlinear mapping, network reasoning and prediction of neural network, but also takes advantage of the global optimization of genetic algorithm to get a reasonable optimization result. Compared with the traditional structural optimization methods, this method has higher efficiency and accuracy good.