NA-PTA-NA Lab5 正则化幂法求矩阵特征值

SydneyCarton_

于 2021-04-15 21:32:04 发布

阅读量348

点赞数

分类专栏：数值分析

本文链接：https://blog.csdn.net/SydneyCarton_/article/details/115740529

版权

该博客介绍了一种使用正则化幂法来近似计算给定方阵A的特征值及其对应特征向量的方法。函数接受矩阵大小、初始特征值近似、初始向量、容差和最大迭代次数作为参数，并在满足精度要求或达到最大迭代次数时返回结果。如果初始值已经非常接近准确的特征值，函数将返回-1，否则返回1表示找到解，0表示超过最大迭代次数。

摘要由CSDN通过智能技术生成

Approximating Eigenvalues
Approximate an eigenvalue and an associated eigenvector of a given n×n matrix A near a given value p and a nonzero vector x =(x1 ,⋯,xn )^T .

Format of function:

int EigenV(int n, double a[][MAX_SIZE], double *lambda, double v[], double TOL, int MAXN);

where int n is the size of the matrix of which the entries are in the array double a[][MAX_SIZE] and MAX_SIZE is a constant defined by the judge; double *lambda is passed into the function as an initial approximation p of the eigenvalue and is supposed to be returned as a more accurate eigenvalue; double v[] is passed into the function as the initial vector x and is supposed to be returned as the associated eigenvector with unit∥⋅∥∞ norm; double TOL is the tolerance for the eigenvalue; and finally int MAXN is the maximum number of iterations.

The function must return:

1 if there is a solution;
0 if maximum number of iterations exceeded;
−1 if p is just the accurate eigenvalue.

Sample program of judge:

#include <stdio.h>

#define MAX_SIZE 10

int EigenV(int n, double a[][MAX_SIZE], double *lambda, double v[], double TOL, int MAXN);

int main()
{
   
    int n, MAXN, m, i, j, k;
    double a[MAX_SIZE][MAX_SIZE], v[MAX_SIZE];
    double lambda, TOL;

    scanf("%d", &n);
    for (i=0; i<n; i++) 
        for (j=0; j<n; j++) 
            scanf("%lf", &a[i][j]);
    scanf("%lf %d", &TOL, &MAXN);
    scanf("%d", &m);
    for (i=0; i<m; i++) {
   
        scanf("%lf", &lambda);
        for (j=0; j<n; j++)
            scanf("%lf", &v[j]);
        switch (EigenV(n, a, &lambda