# Inequality about Eigenvalue of a Real Symmetric Matrix

## Problem 451

Let $A$ be an $n\times n$ real symmetric matrix.

Prove that there exists an eigenvalue $\lambda$ of $A$ such that for any vector $\mathbf{v}\in \R^n$, we have the inequality

\[\mathbf{v}\cdot A\mathbf{v} \leq \lambda \|\mathbf{v}\|^2.\]

Add to solve later

Sponsored Links

## Proof.

Recall that all the eigenvalues of a symmetric matrices are real numbers.

Let $\lambda_1, \dots, \lambda_n$ be eigenvalues of $A$.

Since these eigenvalues are real numbers, there is the largest one.

Let $\lambda$ be the largest eigenvalue of $A$.

With this choice of $\lambda$ we show that the inequality

\[\mathbf{v}\cdot A\mathbf{v} \leq \lambda \|\mathbf{v}\|^2\]
holds for any $\mathbf{v}\in \R^n$.

Also recall that for a real symmetric matrix, there are eigenvalues $\mathbf{v}_1, \dots, \mathbf{v}_n$ corresponding to $\lambda_1, \dots, \lambda_n$ such that

\[B=\{\mathbf{v}_1, \dots, \mathbf{v}_n\}\]
form an orthonormal basis of $\R^n$.

(This statement is equivalent to that every real symmetric matrix is diagonalizable by an orthogonal matrix.)

Let $\mathbf{v}$ be an arbitrary vector in $\R^n$.

Then since $B$ is a basis of $\R^n$, we can write

\[\mathbf{v}=c_1\mathbf{v}_1+\dots+c_n\mathbf{v}_n\] for some $c_1, \dots, c_n\in \R$.

Then we calculate

\begin{align*}

A\mathbf{v}&=A(c_1\mathbf{v}_1+\dots+c_n\mathbf{v}_n)\\

&=c_1A\mathbf{v}_1+\dots+c_nA\mathbf{v}_n\\

&=c_1\lambda_1\mathbf{v}_1+\dots+c_n\lambda_n\mathbf{v}_n

\end{align*}

since $A\mathbf{v}_i=\lambda_i\mathbf{v}_i$ for $i=1, \dots, n$.

Using this, we have

\begin{align*}

\mathbf{v}\cdot A\mathbf{v}&=(c_1\mathbf{v}_1+\dots+c_n\mathbf{v}_n)\cdot (c_1\lambda_1\mathbf{v}_1+\dots+c_n\lambda_n\mathbf{v}_n)\\

&=c_1^2\lambda_1+\cdots+c_n^2\lambda_n.

\end{align*}

Here, we used that $B=\{\mathbf{v}_1, \dots, \mathbf{v}_n\}$ is an orthonormal basis of $\R^3$.

That is, we used the properties

\begin{align*}

\mathbf{v}_i\cdot \mathbf{v}_j=\begin{cases}

1 & \text{if } i=j\\

0 & \text{if } i\neq j.

\end{cases}

\end{align*}

Since $\lambda$ is the largest eigenvalue of $A$, we have

\begin{align*}

\mathbf{v}\cdot A\mathbf{v}&=c_1^2\lambda_1+\cdots+c_n^2\lambda_n\\

& \leq c_1^2\lambda+\cdots+c_n^2\lambda\\

&=\lambda(c_2^2+\cdots+c_n^2)\\

&=\lambda \|\mathbf{v}\|^2.

\end{align*}

Hence the required inequality holds.

Add to solve later

Sponsored Links

## 1 Response

[…] For a proof of this problem, see the post “Inequality about Eigenvalue of a Real Symmetric Matrix“. […]