Discrete Arzel`a-Ascoli Theorem on the Half Line and Its Application
2018-08-17LUYANQIONGANDMARUYUN
LU YAN-QIONG AND MA RU-YUN
(Department of Mathematics,Northwest Normal University,Lanzhou,730070)
Abstract:The compactness criterion inis given and as an application,we study the existence of positive solutions of second order boundary value problems of difference equation on the half line by the fixedpoint theorem in cones.
Key words:discrete Arzel`a-Ascoli theorem,half line,positive solution,difference equation, fixed point
1 Introduction
The well-known Arzel`a-Ascoli theorem gives necessary and sufficient conditions to decide whether every subsequence of a given sequence of real-valued continuous functions de fined on a closed and bounded interval has a uniformly convergent subsequence.The theorem is a fundamental result in functional analysis.In particular,it forms the basis for the proof of the Peano existence theorem in the theory of ordinary differential equations(see[1])and Montel’s theorem in complex analysis(see[2]).It also plays a decisive role in the proof of the Peter-Weyl theorem(see[3]).
First we brie fl y recall some history about Arzel`a-Ascoli theorem.The notion of equicontinuity was introduced at around the same time by Arzel`a[4]and Ascoli[5].A weak form of the theorem was proven by Ascoli[5],who established the sufficient condition for compactness,and by Arzel`a[6],who established the necessary condition and gave the first clear presentation of the result.A further generalization of the theorem was proven by Fr´echet[7]to sets of real-valued continuous functions with domain a compact metric space,see Dunford and Schwartz’s work(see[8]).Modern formulations of the theorem allow for the domain to be compact Hausdorffand for the range to be an arbitrary metric space.More general formulations of the theorem exist that give necessary and sufficient conditions for a family of functions from a compactly generated Hausdorffspace into a uniform space to be compact in the compact-open topology[9].Later,many mathematicians,such as Corduneanu[10],Kelley and Namioka[11],Kelley[9]developed their work.The investigation of compactness criterion has involved much interest in the new century(see[10]–[12]).The compact criterion in

has been investigated in Chapter 2 of[10],the result is as follows:
Theorem ALet F⊂Clbe a set satisfying the following conditions:
(i)F is bounded in Cl;
(ii)the function belonging to F are equicontinuous on any compact interval of[0,∞).
(iii)the function from F are equiconvergent,i.e.,given ε> 0,there corresponds T(ε)> 0 such that‖f(t)− f(∞)‖ < ε for any t≥ T(ε)and f ∈ F.Then F is compact in Cl.
It has been pointed out that Theorem A plays an important role to study the positive solutions of second order boundary value problem of ordinary differential equation on the half line by the fixed-point theorem in cones(see[13]–[15]and their references).However,there is few literature to study the boundary value problems of difference equation on the half line by the fixed-point theorem in cones because of no compact criterion in

A natural question is whether exists the compactness criterion in l⋄?In this paper,we give a positive answer and obtain the necessary and sufficient conditions of the compactness criterion in l⋄.It is particularly useful to study the boundary value problems of difference equation on the half line.
The rest of this paper is organized as follows.In Section 2,we present the discrete Arzel`a-Ascoli Theorem on the half line.Section 3 is devoted to the application of the discrete Arzel`a-Ascoli Theorem.
2 Compactness Results
Let N+={1,2,···},N={0,1,2,···}and[a,b]Z={a,a+1,a+2,···,b}with a < b,a,b∈N+.
Let

with the norm

Then l⋄is a Banach space.
Theorem 2.1Let F ⊂l⋄be a set satisfying the following conditions:
(A1)F is bounded in l⋄;
(A2)the functions from F are equiconvergent,i.e.,given ϵ> 0,there corresponds N(ϵ)>0 such that

Then F is compact in l⋄.
Proof. We sketch the proof of the above criterion.First,from any sequence of F we can extract a subsequence,say{fm},such that{fm(∞)}is convergent.From Condition(A2),it follows that

If m and p are sufficiently large,then

Assume that(2.2)holds for m,p ≥ N(=N(ϵ)).On the other hand,{fm}is uniformly bounded on{0,1,2,···,N}.Therefore,{fm(t)}contains a subsequence that is uniformly convergence on{0,1,2,···,N},say{fmk(t)}.We can write

Now,from(2.2)and(2.3),we say that{fmk(t)}satis fies the Cauchy’s condition on N.Hence,it is uniformly convergent on N,i.e.,it converge in l⋄.
Theorem 2.2Let F ⊂ l⋄be compact in l⋄.Then
(A1)F is bounded in l⋄;
(A2)the functions from F are equiconvergent.Proof. Since F is compact in l⋄,it follows that
(1)F is bounded in l⋄;
(2)for arbitrary ε > 0,there exist f1,···,fk∈ F,such that



Now,for arbitrary f ∈ F,we have from(2.4)that there exists i∈ {1,···,k},such that

In particular,

This together with(2.5),(2.6)imply that for t≥Nj,

Remark 2.1Theorem 2.1 and Theorem 2.2 give the necessary and sufficient conditions of the compactness criterion in l⋄,which can be solved the compactness of summation operator on the in finite dimensional space l⋄.Compared with Theorem A,the condition(ii)of Theorem A is not necessary in the compactness criterion in l⋄.In the following section,Lemmas 3.2 and 3.3 are the speci fic applications of Theorem 2.1.
3 Application
In this section,we investigate the existence of positive solutions for the nonlinear discrete boundary value problem

where k>0 is a constant,f:N+×[0,∞)→[0,∞)is continuous.
First we study some properties of the Green’s function to the linear problem

By a simple computation,Green’s function G(t,s)of(3.2)is given by

where

Useful estimates of the kernel G are provided in the following lemma which the proof is omitted.
Lemma 3.1The Green’s function G(t,s)satis fies the following bounds:
(i)G(t,s)λ−µt≤ G(s,s)λ−s, t,s ∈ N,µ ≥ 1;
(ii)G(t,s) ≥ σG(s,s)λ−s, t ∈ [γ,δ]Z,s ∈ N+(0 < γ < δ),where 0 < σ :=min{λ−δ,λγ−λ−γ}< 1.
Let

endow with the norm

whereµ>1 is a constant.Then X is a Banach space.
As an immediate consequence of Theorem 2.1,we have the following
Lemma 3.2Let D⊂X.If the following conditions hold:
(1)D is bounded in X;
(2)the functions belonging to{y|y(t)=x(t)λ−µt,x ∈ D}are equiconvergent,i.e.,given ϵ> 0,there exists an N(ϵ)> 0 such that

Then D is compact in X.
De fine a cone

De fine the operator A:X→X by

Then x is a solution of problem(3.1)if and only if x is the fixed point of the operator A.
Lemma 3.3Suppose that
(H1)f∈C(N+×[0,+∞),[0,+∞))and there exists p>0 withsuch that

where the functions a,b:N+→[0,∞).
(H2)the summations

are convergent and satisfy

with r>0 is a constant.
Then for any bounded set Ω⊂X,the operator.Moreover,is compact and continuous.
Proof.We divide the proof into three steps.
Choose some bounded set Ω⊂X.The positivity of G,(3.4)and(H1)imply that

By(H1)and Lemma 3.1(i),for any x∈¯Ω∩P,we get the following estimates:

Passing to the supremum over t,we get

Hence


Passing to the supremum over η,we deduce that

This together with(3.5)and(3.7)imply that A(¯Ω∩P)⊂P.

According to the continuous of f and the Lebesgue dominated convergence theorem we have


Next we show that the functions belonging toare equiconvergent.Since(Ax)(∞)=0,we have that for arbitrary ε> 0,there exists an N(ε)>0,such that for any

So

are equiconvergent.
Theorem 3.1Let p > 1 and assume that(H1)–(H2)hold together with

Then problem(3.1)has at least one nontrivial positive solution.
Proof.Let r be de fined by(H2).From(3.8),there exists ansuch that


From Lemma 3.3,it follows thatis completely continuous.
If x∈P∩∂Ω1,then from(H2)and(3.6),it follows that

So

If x∈P∩∂Ω2,then



By Krasnoselskii fixed point theorem(see[16]),the operator A has at least one fixed point inwhich means that the problem(3.1)has a positive solution x such that
The following result deals with the other polynomial growth case and can be proved in a similar method.We omit the proof.
Theorem 3.2Let p < 1 and assume that(H1)–(H2)hold together with

Then problem(3.1)has at least one nontrivial positive solution.
Finally,we give an example to illustrate the results of Theorems 3.1 and 3.2.
Example 3.1Let us consider the boundary value problems


Setµ=2>1.Then the work space is

In addition,from Lemma 3.1,set γ =3,δ=13,it follows that

It follows from f(t,x)=t−2+(3t−3+4)xpyields that(H1)and(H2)holds with

If p=1.5>1,clearly,all conditions of Theorem 3.1 hold,then the problem(3.10)has at least one positive solution.
If p=0.5<1,we also can verify that all conditions of Theorem 3.2 hold,then the problem(3.10)has at least one positive solution.
杂志排行
Communications in Mathematical Research的其它文章
- Exponential Convergence of Finite-dimensional Approximations to Linear Bond-based Peridynamic Boundary Value Problems
- O-convexity of Orlicz-Bochner Spaces with Orlicz Norm
- Embedding Cartesian Product of Some Graphs in Books
- J-clean and Strongly J-clean Rings
- A Note on Homogenization of Parabolic Equation in Perforated Domains
- Existence and Uniqueness of Fixed Points for Mixed Monotone Operators with Application
