Proof of Hölder's Inequality in Lebesgue Spaces
📂Lebesgue SpacesProof of Hölder's Inequality in Lebesgue Spaces
Theorem
Let us consider Ω⊂Rn as an open set. Suppose we’re given two constants 1<p<∞,1<p′<∞ satisfying the following equation.
p1+p′1=1(or p′=p−1p)
If u∈Lp(Ω) and v∈Lp′(Ω), then uv∈L1(Ω), and the inequality below holds.
∥uv∥1=∫Ω∣u(x)v(x)∣dx≤∥u∥p∥v∥p′
This inequality is known as Hölder’s inequality.
Explanation
p′ is known as the Hölder conjugate or conjugate exponent of p. It is often denoted as q.
If ∣u(x)∣p and ∣v(x)∣p′ are proportional almost everywhere in Ω, equality holds.
It is essentially the same as Hölder’s inequality in Euclidean spaces, and it also becomes the Cauchy-Schwarz inequality when p=p′=2. The proof is no different than that of the Cauchy-Schwarz inequality, with the addition of Young’s inequality.
It can also be generalized in the following form.
∥uv∥r=(∫Ω∣u(x)v(x)∣rdx)1/r≤∥u∥p∥v∥p′
∥u∥r=(∫Ω∣u(x)∣rdx)1/r≤j=1∏N∥uj∥pj=∥u1∥p1⋯∥uN∥pN
Proof
Young’s Inequality
Given p1+p′1=1 satisfies and for two constants greater than 1, p,p′, and two positive numbers a,b,
ab≤pap+p′bp′
Case 1. ∥u∥p=0 or ∥v∥p′=0
Since almost everywhere in Ω is u(x)=0 or almost everywhere in Ω is v(x)=0, almost everywhere in Ω is u(x)v(x)=0. Therefore,
∫Ωu(x)v(x)dx=∥uv∥1=0
and
∥u∥p∥v∥p′=0
hence the inequality is satisfied.
Case 2. Other cases
By substituting a=∥u∥p∣u(x)∣ and b=∥v∥p′∣v(x)∣ in Young’s inequality, we have
∥u∥p∣u(x)∣∥v∥p′∣v(x)∣≤p∥u∥pp∣u(x)∣p+p′∥v∥p′p′∣v(x)∣p′
Integrating both sides yields
∥u∥p∥v∥p′1∫Ω∣u(x)v(x)∣dx≤≤≤=p∥u∥pp1∫Ω∣u(x)∣pdx+p′∥v∥p′p′1∫Ω∣v(x)∣p′dxp∥u∥pp1∥u∥pp+p′∥v∥p′p′1∥v∥p′p′p1+p′11
Moving the constant to the left side gives
∥uv∥1=∫Ω∣u(x)v(x)∣dx≤∥u∥p∥v∥p′
Thus, uv∈L1(Ω) holds, and the inequality is satisfied.
■
See also