论文部分内容阅读
O ja连续型全反馈神经网络模型可以有效计算实对称矩阵的主特征向量,该网络的动态行为由描述其模型的微分方程所决定,详细研究了O ja动力系统的稳定性问题.对于非正定实对称矩阵最大特征根为零,且至少有一特征根为负的情形,证明了从单位球外出发的解并不一定必然导致有限逸时,完善了O ja模型计算实对称矩阵主特征向量的收敛性结果,数值实验结果进一步验证了理论分析的正确性.
O ja continuous full feedback neural network model can effectively calculate the main eigenvectors of real symmetric matrices. The dynamic behavior of this network is determined by the differential equation describing its model, and the stability of O ja dynamic system is studied in detail. The fact that the largest eigenvalue of real symmetric matrices is zero and at least one eigenvalue is negative proves that the solution starting from the unit sphere does not necessarily lead to a finite easing and improves the Oja model to calculate the main eigenvectors of real symmetric matrices Convergence results, the numerical experimental results further verify the correctness of the theoretical analysis.