Generalized Sum is Linear

From ProofWiki
Jump to navigation Jump to search

Theorem

Let $\family {z_i}_{i \mathop \in I}$ and $\family {w_i}_{i \mathop \in I}$ be $I$-indexed families of complex numbers.

That is, let $z_i, w_i \in \C$ for all $i \in I$.


Let $\ds \sum \set {z_i: i \in I}$ and $\sum \set {w_i: i \mathop \in I}$ converge to $z, w \in \C$, respectively.


Then:

$(1): \quad \ds \sum \set {z_i + w_i: i \in I}$ converges to $z + w$
$(2): \quad \forall \lambda \in \C: \ds \sum \set {\lambda z_i: i \in I}$ converges to $\lambda z$


Proof

Proof of $(1)$

Let $\epsilon > 0$.

To verify the convergence, it is necessary to find a finite $F \subseteq I$ such that:

$\ds \map d {\sum_{i \mathop\in G} z_i + w_i, z + w} < \epsilon$ for all finite $G$ with $F \subseteq G \subseteq I$


Now let $F_z, F_w \subseteq I$ be finite subsets of $I$ such that:

$\ds \map d {\sum_{i \mathop \in G} z_i, z} < \frac \epsilon 2$ for all finite $G$ with $F_z \subseteq G \subseteq I$
$\ds \map d {\sum_{i \mathop \in G} w_i, w} < \frac \epsilon 2$ for all finite $G$ with $F_w \subseteq G \subseteq I$


The set $F_z \cup F_w$ will be the sought $F$. Let $G$ be finite such that $F_z \cup F_w \subseteq G \subseteq I$.

It follows that:

\(\ds \map d {\sum_{i \mathop \in G} z_i + w_i, z + w}\) \(=\) \(\ds \size {\paren {\sum_{i \mathop \in G} z_i + w_i} - \paren {z + w} }\) Definition of Metric Induced by Norm
\(\ds \) \(=\) \(\ds \size {\paren {\sum_{i \mathop \in G} z_i - z} + \paren {\sum_{i \mathop \in G} w_i - w} }\)
\(\ds \) \(\le\) \(\ds \size {\paren {\sum_{i \mathop \in G} z_i - z} } + \size {\paren {\sum_{i \mathop \in G} w_i, w} }\) Triangle Inequality for $\size {\, \cdot \,}$
\(\ds \) \(=\) \(\ds \map d {\sum_{i \mathop\in G} z_i, z} + \map d {\sum_{i \mathop\in G} w_i, w}\) Definition of Metric Induced by Norm
\(\ds \) \(<\) \(\ds \epsilon\) $F_z, F_w \subseteq G$

From the definition of convergence, it is concluded that:

$\ds \sum \set {z_i + w_i: i \in I} = z + w$

$\Box$


Proof of $(2)$

Let $\epsilon > 0$.

To verify the convergence, it is necessary to find a finite $F \subseteq I$ such that:

$\ds \map d {\sum_{i \mathop\in G} \lambda z_i, \lambda z} < \epsilon$ for all finite $G$ with $F \subseteq G \subseteq I$


Now let $F_z \subseteq I$ be a finite subset of $I$ such that:

$\ds \map d {\sum_{i \mathop \in G} z_i, z} < \frac \epsilon {\size \lambda}$

for all finite $G$ with $F_z \subseteq G \subseteq I$

The set $F_z$ will be the sought $F$.

Let $G$ be finite such that $F_z \subseteq G \subseteq I$.

It follows that:

\(\ds \map d {\sum_{i \mathop \in G} \lambda z_i, \lambda z}\) \(=\) \(\ds \size {\paren {\sum_{i \mathop \in G} \lambda z_i} - \lambda z}\) Definition of Metric Induced by Norm
\(\ds \) \(=\) \(\ds \size {\lambda \paren {\sum_{i \mathop \in G} z_i - z} }\)
\(\ds \) \(=\) \(\ds \size \lambda \size {\sum_{i \mathop \in G} z_i - z}\) Multiplicativity for $\size {\, \cdot \,}$
\(\ds \) \(=\) \(\ds \size \lambda \map d {\sum_{i \mathop \in G} z_i, z}\) Definition of Metric Induced by Norm
\(\ds \) \(<\) \(\ds \epsilon\) $F_z \subseteq G$

From the definition of convergence, conclude that:

$\ds \sum \set {\lambda z_i: i \in I} = \lambda z$

$\blacksquare$


Also see