Grocery on Distributed Algorithms

T1: A Simple Question 2: Parseval's Inequality

Reminder: This post contains 591 words · 2 min read · by Xianbin

Parseval’s Inequality

Let \(A = \{u_1,\ldots,u_n\}\) be an orthonormal basis for an inner-product space \(S\), and let \(\textbf{x} = \sum_i \xi_i u_i\) be the Fourier expansion of \(\textbf{x} \in S\). Prove that

\[\sum_{i=1}^n \lvert \xi_i\rvert^2 = \lVert \textbf{x} \rVert^2\]

\(\textbf{Proof}\).

\[\begin{align} \lVert \textbf{x} \rVert^2 = \lVert \xi_1u_1+\ldots+\xi_n u_n\lVert^2 \\ =\sum_i^n\lVert \xi_i u_i\rVert^2 \\= \sum_i^n \lvert \xi_i \rvert \end{align}\]

Now, let us prove Equality 2.

\[\lVert \xi_1u_1+\ldots+\xi_n u_n\lVert^2 =\sum_{i=1}^n (\xi_i u_i)\cdot (\xi u_i)^\intercal + \sum_i \Pi_{i\neq j} \xi_i \xi_j u_i u_j = \sum_i^n\lVert \xi_i u_i\rVert^2\]