# Prove $g_n f_n \rightarrow gf$ in $L_p([0,1])$. Where $f_n$ converges and $(g_n)_{n=1}^\infty$ is a sequence of bounded measurable functions

Suppose that $f_n \rightarrow f$ in $L_p([0,1])$ with respect to the $|| . ||_p$ norm. $\{g_n\}_{n=1}^{\infty}$is a sequence of measurable functions such that $|g_n| \leq M$ for each $n$ and $g_n \rightarrow g$.

Work done so far:

$g_n \rightarrow g$ a.e. implies that $g_n^p \rightarrow g^p$ a.e.

Next since $|g_n|^p \leq M^p \in \mathbb{R}$ by Lebesgue Dominated Convergence Theorem we get $\int_{[0,1]} g_n^p \rightarrow \int_{[0,1]} g^p.$

Then, $\int_{[0,1]} |g_n f_n - gf|^p \leq \int_{[0,1]} |g_nf_n - g_nf|^p + \int_{[0,1]} |g_n f - gf|^p$ (by Minkowski's)

$= \int_{[0,1] }|g_n|^p|f_n-f|^p + \int_{[0,1]} |f|^p |g_n - g|^p$

So at this point I will put $(\int_{[0,1]} |g_n f_n - gf|^p)^{1/p} = || g_n f_n - gf||$

And I can almost show that the last two terms go to 0, but am having trouble dealing with $|g_n|^p$ in the first term and $|f|^p$ in the second term.

• Does $|f|^P$ really matter in the limit? – ZaighumRajput Aug 2 '13 at 23:40
• Your inequality "by Minkowski" needs $p$'th roots... – David Mitra Aug 3 '13 at 0:12
• Sure, I can do this: $|| g_nf_n - gf || \leq || g_n - gnf|| + ||g_n f - gf||$ if I just take everything to the $1/p$. Right? – ZaighumRajput Aug 3 '13 at 0:14
• $|| .. ||_p$ * – ZaighumRajput Aug 3 '13 at 0:20

For the first term, use the uniform boundedness of $(g_n)$ to write $$\Bigr(\int |g_n|^p|f_n-f|^p\Bigl)^{1/p}\le M\Bigl(\int |f_n-f|^p\Bigl)^{1/p}$$ and use the fact that $(f_n)$ converges to $f$ in $L_p$.
1. Since $f\in L_p$ and since $(g_n)$ is uniformly bounded, for any $\epsilon>0$, there is a $\delta>0$ so that for any set $A$ of measure less than $\delta$, we have $\int_A |f|^p |g_n-g|^p <\epsilon$.
2. For any $\delta>0$, there is a set $B$ of measure less than $\delta$ so that $(g_n)$ converges uniformly to $g$ on $B^c$. Note for such a $B$, on $B^c$ the quantity $|g_n-g|^p$ can be made uniformly small provided $n$ is sufficiently large.
$$\Bigl(\int |f|^p |g_n-g|^p\Bigr)^{1/p} =\Bigl(\int_B |f|^p |g_n-g|^p \,+\,\int_{B^c} |f|^p|g_n-g|^p\Bigr)^{1/p}.$$