Convergence of Square of Linear Combination of Sequences whose Squares Converge

From ProofWiki
Jump to navigation Jump to search

Theorem

Let $\left\langle{x_i}\right\rangle$ and $\left\langle{y_i}\right\rangle$ be real sequences such that the series $\displaystyle \sum_{i \mathop \ge 0} x_i^2$ and $\displaystyle \sum_{i \mathop \ge 0} y_i^2$ are convergent.

Let $\lambda, \mu \in \R$ be real numbers.


Then $\displaystyle \sum_{i \mathop \ge 0} \left({\lambda x_i + \mu y_i}\right)^2$ is convergent.


Proof

Let $n \in \N$.

Then:

$\displaystyle \sum_{i \mathop = 1}^n \left({\lambda x_i + \mu y_i}\right)^2 = \lambda^2 \sum_{i \mathop = 1}^n x_i^2 + \mu^2 \sum_{i \mathop = 1}^n y_i^2 + 2 \lambda \mu \sum_{i \mathop = 1}^n x_i y_i$

By Cauchy's Inequality:

$\displaystyle \sum_{i \mathop = 1}^n x_i y_i \le \left({\sum_{i \mathop = 1}^n x_i^2}\right)^{\frac 1 2} \left({\sum_{i \mathop = 1}^n y_i^2}\right)^{\frac 1 2}$

Hence:

\(\displaystyle \sum_{i \mathop = 1}^n \left({\lambda x_i + \mu y_i}\right)^2\) \(\le\) \(\displaystyle \lambda^2 \sum_{i \mathop = 1}^n x_i^2 + \mu^2 \sum_{i \mathop = 1}^n y_i^2 + 2 \left\vert{\lambda \mu}\right\vert \sum_{i \mathop = 1}^n x_i y_i\)
\(\displaystyle \) \(\le\) \(\displaystyle \lambda^2 \sum_{i \mathop = 1}^n x_i^2 + \mu^2 \sum_{i \mathop = 1}^n y_i^2 + 2 \left\vert{\lambda \mu}\right\vert \left({\sum_{i \mathop = 1}^n x_i^2}\right)^{\frac 1 2} \left({\sum_{i \mathop = 1}^n y_i^2}\right)^{\frac 1 2}\)
\(\displaystyle \) \(=\) \(\displaystyle \left({\left\vert{\lambda}\right\vert \left({\sum_{i \mathop = 1}^n x_i^2}\right)^{\frac 1 2} + \left\vert{\mu}\right\vert \left({\sum_{i \mathop = 1}^n y_i^2}\right)^{\frac 1 2} }\right)^2\)
\(\displaystyle \) \(\le\) \(\displaystyle \left({\left\vert{\lambda}\right\vert \left({\sum_{i \mathop \ge 0} x_i^2}\right)^{\frac 1 2} + \left\vert{\mu}\right\vert \left({\sum_{i \mathop \ge 0} y_i^2}\right)^{\frac 1 2} }\right)^2\)

Thus the sequence of partial sums $\displaystyle \sum_{i \mathop = 1}^n \left({\lambda x_i + \mu y_i}\right)^2$ is bounded above.

We also have that $\displaystyle \sum_{i \mathop = 1}^n \left({\lambda x_i + \mu y_i}\right)^2$ is also increasing.

So by the Monotone Convergence Theorem, $\displaystyle \sum_{i \mathop = 1}^n \left({\lambda x_i + \mu y_i}\right)^2$ is convergent.

$\blacksquare$


Sources