【摘要】 广义特征值问题是有限元动力学计算的基本问题之一。其算法的有效性对大型复杂结构的动力学分析至关重要。从能量的观点出发 ,将神经网络的优化计算能力与Rayleigh极小值原理相结合得到了求解广义特征值问题的极小值及其对应特征向量的神经网络解法 ,为神经网络优化计算应用于有限元结构动力分析打下了一个良好的基础更多还原

【Abstract】 Generalized eigenvalue is one of the basic problems in the finite element dynamic calculation. The validity of the algorithm is essential to the dynamic analysis of large-scale complex structures. In view of energy, this paper obtains the minimized value to solve the problem of finite eigenvalue and the neural network solution if its corresponding eigenvector by integrating the optimized calculating capability and Rayleigh minimized value theory, which lays a solid foundation for the neural network optimization in the finite element structural dynamic analysis.更多还原