東京大学 新領域創成科学研究科 メディカル情報生命専攻 2018年8月実施 問題8
Author
Description
If an
(1) Show the diagonal elements
(2) Show the eigenvalues of real symmetric matrix
(3) Show the eigenvalues of positive definite matrix
(4) Let
(5) Suppose the eigenvectors of positive definite matrix
如果一个
(1) 证明实正定矩阵
(2) 证明实对称矩阵
(3) 证明正定矩阵
(4) 设
(5) 假设正定矩阵
Kai
(1)
Consider the standard basis vector
Since
Thus,
(2)
For a real symmetric matrix
where
for any vectors
Since
(3)
Let
Taking the transpose of
Since
(4)
Definitions and Setup
- Let
be an real symmetric positive definite matrix. - The set
represents the set of unit vectors in . - We aim to show that the maximum value of the quadratic form
over the set is the largest eigenvalue of .
Step-by-Step Proof
- Spectral Theorem Application:
Since
is a real symmetric matrix, by the spectral theorem, it can be diagonalized as:
where
- Quadratic Form Transformation:
For any unit vector
(i.e., ), we can express in terms of the orthonormal basis formed by the columns of :
where
- Maximization over Unit Vectors:
The expression
can be written as:
where
-
Eigenvalue Maximization:
To maximize
, note that the maximum value occurs when and (because is the largest eigenvalue). Hence,
Therefore,
- Conclusion:
The maximum value of the quadratic form
over the unit sphere is indeed the largest eigenvalue of the positive definite matrix .
(5)
Given:
is a positive definite matrix. has distinct eigenvalues . - The largest eigenvalue
and its associated eigenvector are known.
We aim to find the second largest eigenvalue
Orthogonal Projection Method
-
Orthogonality of Eigenvectors: Since
is symmetric, its eigenvectors corresponding to distinct eigenvalues are orthogonal. Thus, is orthogonal to all other eigenvectors of . -
Defining the Subspace Orthogonal to
: Let be the subspace of consisting of vectors orthogonal to :
- Rayleigh Quotient in the Subspace:
The Rayleigh quotient
for a vector is given by:
We need to maximize this quotient over the subspace
-
Projection Method: To find
, we consider the effect of and deflate by projecting it onto the subspace orthogonal to . Define the matrix
as:
This matrix
-
Finding
: The second largest eigenvalue of will be the largest eigenvalue of in the subspace orthogonal to . To find
, consider any vector in :
Here,
- Maximizing the Rayleigh Quotient:
The Rayleigh quotient in the subspace
for the matrix becomes:
Since
Therefore, the maximum value of
- Conclusion:
By maximizing the Rayleigh quotient in the subspace orthogonal to the eigenvector
associated with , we find the second largest eigenvalue .
Knowledge
正定矩阵 特征值和特征向量 Rayleigh商 正交性
难点思路
计算次大特征值的过程可能是一个难点,尤其是如何正确地使用正交性和 Rayleigh 商。
解题技巧和信息
- 正定矩阵的定义和性质非常重要,尤其是对特征值的影响。
- 计算特征值时,Rayleigh 商是一个有效工具。
- 正交性在分解和简化问题中非常有用。
重点词汇
- Positive definite matrix 正定矩阵
- Eigenvalue 特征值
- Eigenvector 特征向量
- Rayleigh quotient Rayleigh 商
- Orthogonal 正交的
参考资料
- Linear Algebra and Its Applications by Gilbert Strang, Chapter 6.
- Introduction to Linear Algebra by Gilbert Strang, Chapter 7.