Monotone Convergence Theorem

Theorem (Limit of Measurable Functions is Measurable)

Let (X,F,μ)(X,\mathcal{F},\mu) be a measure space. Let (fn)nN(f_{n})_{n\in\mathbb{N}} be a sequence of measurable functions, fn:XRf_{n}:X\to \mathbb{R}. Let f:XRf:X\to \mathbb{R} be such that fnff_{n}\to f pointwise, then f:XR is measurablef:X\to \mathbb{R}\text{ is measurable}

Theorem (1.26)

Let (fn)nN(f_{n})_{n\in\mathbb{N}} be a sequence of measurable functions, fn:XR+f_{n}:X\to \mathbb{R}^{+} in measure space (X,F,μ)(X,\mathcal{F},\mu). Assume fnff_{n}\uparrow f pointwise i.e. 0f0f1fnf0\le f_{0}\le f_{1}\le\dots\le f_{n}\le\dots fthen, limnXfndμ=Xfdμ\lim_{ n \to \infty } \int\limits _{X}f_{n} \, d\mu=\int\limits _{X}f \, d\mu

\begin{proof} First we note using the fact that that ff is measurable.

Then, since 0fnfn+10\le f_{n}\le f_{n+1} n\forall n, then 0fndμfn+1dμ0\le\int\limits f_{n} \, d\mu\le\int\limits f_{n+1} \, d\mu since ν\nu is a measure. Hence, (fndμ)nN\left( \int\limits f_{n} \, d\mu \right)_{n\in\mathbb{N}} has a limit in Rˉ\bar{\mathbb{R}}.

Now let α=limn(fndμ)Rˉ\alpha=\lim_{ n \to \infty }\left( \int\limits f_{n} \, d\mu \right)\in \bar{\mathbb{R}}, we WTS fdμ=α\int\limits f \, d\mu =\alpha


Since n1\forall n\ge 1, 0fnf0\le f_{n}\le f then 0fndμfdμ0limnfndμfdμ0αfdμ\begin{gather} 0\le &\int\limits f_{n} \, d\mu &\le &\int\limits f \, d\mu \\ 0\le &\lim_{ n \to \infty } \int\limits f_{n} \, d\mu &\le &\int\limits f \, d\mu\\ 0\le &\alpha &\le &\int\limits f \, d\mu \end{gather}Now TS αfdμ\alpha\ge \int\limits f \, d\mu we instead show 0<c<1:cfdμα\forall 0<c<1:c\int\limits f \, d\mu \le \alpha \end{proof}

Theorem (Countable Additivity is Necessary for MCT)

Countable additivity is a necessary condition for Monotone Convergence Theorem.

Linked from