WU Di, DENG Yangkendi, YU Dandan, YAN Dunyan?
(1 School of Science, Zhejiang University of Science and Technology,Hangzhou 310023, China;2 School of Mathematical Sciences, University of Chinese Academy of Sciences,Beijing 100049, China)
Abstract In this paper, we give a novel proof for the following equality
for f∈Lp,q(X,μ) with 0
Keywords limiting behavior; distribution functions; Lorentz spaces
Letfbe a measurable function on a measure space (X,μ) and 0
(1)
wheredf(α)=μ{x:|f(x)|>α} is the distribution function off. The set of allfwith |f|Lp,q<∞ is denoted byLp,q(X,μ) and is called the Lorentz space with indicespandq.
There are many simple properties of Lorentz space. For 0
Lp,p(X,μ)=Lp(X,μ),
and
Lp,∞(X,μ)=WLp(X,μ).
For 0
‖f‖Lp,r≤cp,q,r‖f‖Lp,q.
More properties of Lorentz space can be found in Refs.[1-4].
The limiting property of distribution functions ofLpfunctions has been proved in Ref.
[5] for 1≤p<∞. That is, whenf∈Lp,p(n), we have that
(2)
However, forf∈L1(n), we know thatMf∈L1,∞(n), whereMis the maximal operator. In Ref.
[6], P. Janakiraman has proved that
This means that the limiting equality of distribution functions does not hold forL1,∞function.
It is well-known that the following inclusion relation
Lp,p(n)Lp,q(n)Lp,∞(n)
holds for 0
It is worth investigating the limiting property of distribution functions off∈Lp,q(X,μ).
Now we formulate our main theorem.
Theorem1.1Letμbe aσ-finite positive measure on someσ-algebra in setX. Forf∈Lp,q(X,μ) with 0
df(α):=μ({x∈X:|f(x)|>α}).
Then we have
(3)
ProofFirst, we prove that
If the conclusion does not hold, then we conclude that
or
and
forn∈.Obviously, there exists a strictly decreasing subsequence, still denoted as {xn}.
Sincedfis a decreasing function, it follows that
(4)
Settingλn=(xn+1/xn)q∈(0,1), the basic principle of mathematics analysis tells us that the infinite product ∏λnconverges to a nonzero number if and only if ∑(1-λn) converges. However we deduce that
It implies from (4) that
‖f‖Lp,q=∞.
This leads to contradiction.
Next, we prove
such that
and
Therefore, we conclude that
(5)
Note that the infinite product ∏(xn/xn+1)qdiverges to 0. Forα>0,αpdf(α)≥0 implies
That leads to (3).
□
Now, we begin to prove that the functionαpin (2) can not be improved. For this purpose, we first give the following lemma which characterize the special property of some function.
Lemma1.1Letg:(0,+∞)→[0,+∞) be a right-continuous non-increasing function. Then there exists a functionf:n→[0,+∞] such thatdf(α)=g(α) for anyα∈(0,+∞).
ProofLet
(6)
wherevn=m(B(0,1)) andB(x,r) denotes a ball with the center atxand the radiusr.
Then we merely need to prove that
(7)
holds, for anyα∈(0,+∞).
Assume
Note thatgis a non-increasing function. We conclude from definition offin (6) that
≤α.
(8)
Thus the inequality (8) implies that the set of the left hand of (7) is contained in the right hand.
Assume
Note thatgis a right-continuous function. Then there existsα′>αsuch that
Then we conclude that
=α′>α.
This means that
In a word, (7) holds.
□
As an application of Lemma 1.2, we prove that the functionαpcan not be improved for some sense.
Theorem1.2Suppose 0
(9)
there exists a functionf∈Lp,q(n) such that
h(αk)≥4k
and
xk+1 fork∈andαk∈[2-1xk,xk). Define a functiong:(0,+∞)→[0,+∞) as (10) By Lemma 1.1, there exists a functionf:n→[0,+∞] such thatdf(α)=g(α) for anyα∈(0,+∞).And we have from the definition ofgin (10) that This impliesf∈Lp,q(n). On the other hand, it follows that Consequently, we obtain □ RemarkIn the fact, we can also prove that, for anyhsatisfying there exists a functionf∈Lp,q(n) such that where 0 Theorem 1.2 tells us that the functionαpin (3) can not be improved. Forh(α)=|logα|swheres>0, we have some more explicit conclusions. Corollary1.1Supposesis a positive constant and 0 p/q, then there exists a functionf∈Lp,q(n) such that if 0 (11) ProofFirst, we assume thats>p/q.Letg:(0,+∞)→[0,+∞) be defined by whereais a sufficiently small positive constant such thatg(α) is non-increasing. By Lemma 1.1, there exists a functionf:n→[0,+∞] such thatdf(α)=g(α) for anyα∈(0,+∞). In fact, we can choose the functionfas in (6). It implies from simple calculation that f∈Lp,q(n) and Next, we assume that 0 or