# Equivalence of Definitions of Variation of Signed Measure

## Theorem

The following definitions of the concept of Variation of Signed Measure are equivalent:

### Definition 1

Let $\struct {X, \Sigma}$ be a measurable space.

Let $\mu$ be a signed measure on $\struct {X, \Sigma}$.

Let $\struct {\mu^+, \mu^-}$ be the Jordan decomposition of $\mu$.

We define the variation $\size \mu$ of $\mu$ by:

$\size \mu = \mu^+ + \mu^-$

### Definition 2

Let $\struct {X, \Sigma}$ be a measurable space.

Let $\mu$ be a signed measure on $\struct {X, \Sigma}$.

Let $\map P A$ be the set of finite partitions of $A$ into $\Sigma$-measurable sets.

We define the variation $\cmod \mu : \Sigma \to \overline \R$ of $\mu$ by:

$\ds \map {\cmod \mu} A = \sup \set {\sum_{j \mathop = 1}^n \cmod {\map \mu {A_j} } : \set {A_1, A_2, \ldots, A_n} \in \map P A}$

for each $A \in \Sigma$, where the supremum is taken in the set of extended real numbers $\overline \R$.

## Proof

We aim to prove that for each $A \in \Sigma$, we have:

$\ds \map {\mu^+} A + \map {\mu^-} A = \sup \set {\sum_{j \mathop = 1}^n \cmod {\map \mu {A_j} } : \set {A_1, A_2, \ldots, A_n} \in \map P A}$

Let $\tuple {P, N}$ be a Hahn decomposition for $\mu$.

Then $\set {P \cap A, N \cap A}$ is a finite partition of $A$ into $\Sigma$-measurable sets.

Then, we have from the definition of Jordan decomposition:

$\map \mu {P \cap A} = \map {\mu^+} A \ge 0$

and:

$\map \mu {N \cap A} = -\map {\mu^-} A \le 0$

Then:

$\cmod {\map \mu {P \cap A} } = \map {\mu^+} A$

and:

$\cmod {\map \mu {N \cap A} } = \map {\mu^-} A$

so:

 $\ds \map {\mu^+} A + \map {\mu^-} A$ $=$ $\ds \size {\map \mu {P \cap A} } + \size {\map \mu {N \cap A} }$ $\ds$ $\le$ $\ds \sup \set {\sum_{j \mathop = 1}^n \cmod {\map \mu {A_j} } : \set {A_1, A_2, \ldots, A_n} \in \map P A}$ Definition of Supremum of Subset of Extended Real Numbers

Conversely, for each $\set {A_1, A_2, \ldots, A_n} \in \map P A$, we have:

 $\ds \sum_{j \mathop = 1}^n \size {\map \mu {A_j} }$ $=$ $\ds \sum_{j \mathop = 1}^n \size {\map {\mu^+} {A_j} - \map {\mu^-} {A_j} }$ Definition of Jordan Decomposition $\ds$ $\le$ $\ds \sum_{j \mathop = 1}^n \paren {\size {\map {\mu^+} {A_j} } + \size {\map {\mu^-} {A_j} } }$ Triangle Inequality $\ds$ $=$ $\ds \sum_{j \mathop = 1}^n \paren {\map {\mu^+} {A_j} + \map {\mu^-} {A_j} }$ since $\mu^+ \ge 0$ and $\mu^- \ge 0$ $\ds$ $=$ $\ds \sum_{j \mathop = 1}^n \map {\mu^+} {A_j} + \sum_{j \mathop = 1}^n \map {\mu^-} {A_j}$ $\ds$ $=$ $\ds \map {\mu^+} A + \map {\mu^-} A$ using finite additivity of $\mu^+$ and $\mu^-$

so:

$\ds \sup \set {\sum_{j \mathop = 1}^n \cmod {\map \mu {A_j} } : \set {A_1, A_2, \ldots, A_n} \in \map P A} \le \map {\mu^+} A + \map {\mu^-} A$

so:

$\ds \map {\mu^+} A + \map {\mu^-} A = \sup \set {\sum_{j \mathop = 1}^n \cmod {\map \mu {A_j} } : \set {A_1, A_2, \ldots, A_n} \in \map P A}$

$\blacksquare$