Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $\{f_n\}$ and $\{g_n\}$ be sequences of uniformly integrable functions on $E$. Show that for $\alpha$, $\beta$, the sequences $\alpha f_n + \beta g_n$ are also uniformly integrable.

Attempt at a proof:

Since both sequences are uniformly integrable I can find a $\delta=\min\{\delta_{g_n}, \delta_{f_n}\}$ such that $|\int_{A_{f_n}\cup A_{g_n}} f_n +g_n|< \epsilon$.

I would like to know how to show this if I knew that $f_n$ would converge pointwise to $f$. I'm not given this detail however.

Edit:

A family $\mathscr{F}$ of measurable functions on $E$ is said to be uniformly integrable over $E$ provided for each $\epsilon >0$, there is a $\delta >0$ such that for each $f \in \mathscr{F}$, if $A \subseteq E$ is measurable and $m(A)<\delta$, the $\int_A |f|< \epsilon$.

share|improve this question

1 Answer 1

The solution depends on the definition of uniform integrability you use. The most convenient one states that a collection $H$ of integrable functions bounded in $L^1(\Omega,\mathcal F,\mu)$ is uniformly integrable if $\int\limits_A|h|\to0$ when $\mu(A)\to0$, uniformly on $h$ in $H$, that is, $$ \forall\varepsilon\gt0,\ \exists\eta\gt0,\ \forall A\in\mathcal F,\ \forall h\in H,\ \mu(A)\leqslant\eta\implies\int_A|h|\mathrm d\mu\leqslant\varepsilon. $$ Then the proof is direct. First, if $\alpha H=\{\alpha h\mid h\in H\}$ and $H$ is uniformly integrable, then $\alpha H$ is bounded in $L^1$ and, if $\eta(\varepsilon,H)$ makes the implication above true for $H$ and $\varepsilon$, for each $\varepsilon$, then $\eta\left(\frac{\varepsilon}{|\alpha|},H\right)$ makes it true for $\alpha H$ and $\varepsilon$.

Second, if $H+K=\{h+k\mid h\in H+K\}$ and $H$ and $K$ are uniformly integrable, then $H+K$ is bounded in $L^1$ and $\min\{\eta(\varepsilon,H),\eta(\varepsilon,K)\}$ makes the implication above true for $H+K$ and $2\varepsilon$.

Hence for every uniformly integrable collections $H$ and $K$ and every scalar $\alpha$ and $\beta$, $\alpha H+\beta K$ is uniformly integrable as well.

Note finally that the sequences $(f_n)_{n\in\mathbb N}$ and $(g_n)_{n\in\mathbb N}$ are uniformly integrable if and only if the collections $H=\{f_n\mid n\in\mathbb N\}$ and $K=\{g_n\mid n\in\mathbb N\}$ are uniformly integrable and that, then, $\alpha H+\beta K$ contains every function $\alpha f_n+\beta g_n$ (and many more). Hence the uniform integrability of $\alpha H+\beta K$ implies the one of $\{\alpha f_n+\beta g_n\mid n\in\mathbb N\}$.


If one wishes to use the definition that $H$ is uniformly integrable if and only if $\int\limits_{|h|\gt t}|h|\mathrm d\mu\to0$ when $t\to+\infty$, uniformly on $h$ in $H$, that is, $$ \forall\varepsilon\gt0,\ \exists t,\ \forall s,\ \forall h\in H,\ s\geqslant t\implies\int_{|h|\geqslant s}|h|\mathrm d\mu\leqslant\varepsilon, $$ then the proof is similar but perhaps less convenient. Denote by $t(\varepsilon,H)$ any value of $t$ making the implication above true for $H$ and $\varepsilon$. Then $t(\varepsilon,H)$ works for $\alpha H$ and $|\alpha|\varepsilon$ hence $H$ uniformly integrable implies $\alpha H$ uniformly integrable. For the sum $H+K$, one can convince oneself of the validity of the pointwise inequality $$ |h+k|\,\mathbf 1_{|h+k|\geqslant 2t}\leqslant 2|h|\,\mathbf 1_{|h|\geqslant t}+2|k|\,\mathbf 1_{|k|\geqslant t}, $$ and deduce that, in our notations, $\max\{t(\varepsilon,H),t(\varepsilon,K)\}$ works for $H+K$ and $4\varepsilon$. QED.

share|improve this answer
    
I added my definition for uniform integrability. –  Mark Nov 18 '12 at 23:09
    
OK. Beware though, that there is a subtlety which makes your definition incorrect when $\mu$ has some atoms. Assume for example that no nonempty set $A$ is such that $\mu(A)\lt\delta$... :-) –  Did Nov 18 '12 at 23:16
    
Atoms? My definition is from Royden. –  Mark Nov 18 '12 at 23:22
    
Atoms. –  Did Nov 18 '12 at 23:36

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.