doi:10.1155/2009/496135
Research Article
Sturm-Picone Comparison Theorem of
Second-Order Linear Equations on Time Scales
Chao Zhang and Shurong Sun
School of Science, University of Jinan, Jinan, Shandong 250022, China
Correspondence should be addressed to Chao Zhang,ss [email protected]
Received 29 December 2008; Revised 13 March 2009; Accepted 28 May 2009
Recommended by Alberto Cabada
This paper studies Sturm-Picone comparison theorem of second-order linear equations on time scales. We first establish Picone identity on time scales and obtain our main result by using it. Also, our result unifies the existing ones of second-order differential and difference equations.
Copyrightq2009 C. Zhang and S. Sun. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
1. Introduction
In this paper, we consider the following second-order linear equations:
p1txΔt
Δ
q1txσt 0, 1.1
p2tyΔt
Δ
q2tyσt 0, 1.2
where t ∈ α, β ∩T, pΔ1t, pΔ2t, q1t, and q2t are real and rd-continuous functions in
α, β ∩T.Let Tbe a time scale, σtbe the forward jump operator in T,yΔ be the delta derivative, andyσt:yσt.
First we briefly recall some existing results about differential and difference equations. As we well know, in 1909, Picone1 established the following identity.
Picone Identity
Ifxtandytare the nontrivial solutions of
p1txtq1txt 0,
wheret ∈ α, β , p1t, p2t, q1t,andq2tare real and continuous functions inα, β .If
yt/0 fort∈α, β ,then
xt
yt
p1txtyt−p2tytxt
p1t−p2tx2t q2t−q1tx2t p2t
xtyt
yt −xt
2 .
1.4
By 1.4, one can easily obtain the Sturm comparison theorem of second-order linear differential equations1.3.
Sturm-Picone Comparison Theorem
Assume thatxtandytare the nontrivial solutions of1.3anda, bare two consecutive zeros ofxt,if
p1t≥p2t>0, q2t≥q1t, t∈a, b , 1.5
thenythas at least one zero ona, b .
Later, many mathematicians, such as Kamke, Leighton, and Reid2–5 developed thier work. The investigation of Sturm comparison theorem has involved much interest in the new century6,7 . The Sturm comparison theorem of second-order difference equations
Δp1t−1Δxt−1 q1txt 0,
Δp2t−1Δyt−1 q2tyt 0,
1.6
has been investigated in8, Chapter 8 , wherep1t≥p2t>0 onα, β1 , q2t≥q1ton
α1, β1 , α, βare integers, andΔis the forward difference operator:Δxt xt1−xt. In 1995, Zhang9 extended this result. But we will remark that in8, Chapter 8 the authors employed the Riccati equation and a positive definite quadratic functional in their proof. Recently, the Sturm comparison theorem on time scales has received a lot of attentions. In
10, Chapter 4 , the mathematicians studied
p1txΔt∇q1txt 0,
p2tyΔt
∇
q2tyt 0,
1.7
wherep1t≥ p2t> 0 andq2t≥ q1tfort ∈ρα, σβ ∩T, y∇ is the nabla derivative,
and they get the Sturm comparison theorem. We will make use of Picone identity on time scales to prove the Sturm-Picone comparison theorem of1.1and1.2.
2. Preliminaries
In this section, some basic concepts and some fundamental results on time scales are introduced.
Let T ⊂ R be a nonempty closed subset. Define the forward and backward jump operatorsσ, ρ:T → Tby
σt inf{s∈T:s > t}, ρt sup{s∈T:s < t}, 2.1
where inf∅ supT, sup∅ infT. A pointt ∈ Tis called right-scattered, right-dense, left-scattered, and left-dense ifσt> t, σt t, ρt< t, andρt t,respectively. We putTkT ifTis unbounded above andTkT\ρmaxT,maxT otherwise. The graininess functions
ν, μ:T → 0,∞are defined by
μt σt−t, νt t−ρt. 2.2
Letfbe a function defined onT.fis said to bedeltadifferentiable att∈Tkprovided there exists a constantasuch that for anyε >0, there is a neighborhoodUofti.e.,U t−δ, tδ∩T for someδ >0with
fσt−fs−aσt−s≤ε|σt−s|, ∀s∈U. 2.3
In this case, denotefΔt :a. Iff isdeltadifferentiable for everyt∈Tk, thenfis said to bedeltadifferentiable onT. Iffis differentiable att∈Tk, then
fΔt
⎧ ⎪ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎪ ⎩
lim s→t s∈T
ft−fs
t−s , ifμt 0, fσt−ft
μt , ifμt>0.
2.4
IfFΔt ftfor allt∈Tk, thenFtis called an antiderivative offonT. In this case, define the delta integral by
t
s
fτΔτ Ft−Fs ∀s, t∈T. 2.5
Moreover, a functionf defined onTis said to be rd-continuous if it is continuous at every right-dense point inTand its left-sided limit exists at every left-dense point inT.
Lemma 2.1. Letf, g:T → Randt∈Tk.
iIffis differentiable att, thenfis continuous att.
iiIffandgare differentiable att, thenfg is differentiable attand
fgΔt fσtgΔt fΔtgt fΔtgσt ftgΔt. 2.6
iiiIffandgare differentiable att, andftfσt/0, thenf−1gis differentiable attand
gf−1Δt gΔtft−gtfΔtfσtft−1.
2.7
ivIffis rd-continuous onT, then it has an antiderivative onT.
Definition 2.2. A functionf:T → Ris said to be right-increasing att0∈T\{maxT}provided
ifσt0> ft0in the case thatt0is right-scattered;
iithere is a neighborhoodUoft0such thatft> ft0for allt∈Uwitht > t0in the
case thatt0is right-dense.
If the inequalities forfare reversed iniandii,fis said to be right-decreasing att0.
The following result can be directly derived from2.4.
Lemma 2.3. Assume thatf :T → Ris differentiable att0 ∈T\ {maxT}.IffΔt0>0,thenf is
right-increasing att0; and iffΔt0<0, thenfis right-decreasing att0.
Definition 2.4. One says that a solutionxtof1.1has a generalized zero attifxt 0 or, iftis right-scattered andxtxσt<0.Especially, ifxtxσt<0,then we sayxthas a node attσt/2.
A functionp:T → Ris called regressive if
1μtpt/0, ∀t∈T. 2.8
Hilger14 showed that fort0 ∈ Tand rd-continuous and regressivep, the solution of the
initial value problem
yΔt ptyt, yt0 1 2.9
is given byep·, t0, where
ept, s exp
t
sξμτ
pτΔτ
withξhz
⎧ ⎪ ⎨ ⎪ ⎩
Log1hz
h , if h /0 z, if h0.
2.10
3. Main Results
In this section, we give and prove the main results of this paper.
First, we will show that the following second-order linear equation:
xΔΔt a
Proof. We first divide the left part of3.4into two parts
From1.1and the product ruleLemma 2.1ii, we have
Theorem 3.3 Sturm-Picone Comparison Theorem. Suppose that xt and yt are the integrating it fromatobwe get
which is a contradiction. Therefore, in Case1,ythas at least one generalized zero ona, b ∩
T.
Case 2. Supposeais a zero ofxt,bσb/2 is a node ofxt, xb<0,andxσb>0.It follows from the assumption thatythas no generalized zeros ona, b ∩Tand thatyt>0 for allt ∈a, b ∩Tthatyσb >0.Hence by2.4andp2t≥ p1t >0 ona, b ∩T, we
have
xb
yb
p1bxΔbyb−p2byΔbxb
yxbbμ1bp1bxσbyb−p2byσbxb
p2b−p1b
xbyb
<0.
3.12
By integration, it follows from3.12andxa 0 that
b
a
xt
yt
p1txΔtyt−p2tyΔtxt
Δ
Δt
xt
yt
p1txΔtyt−p2tyΔtxtb a
xybbp1bxΔbyb−p2byΔbxb
<0.
3.13
So, from3.9and above argument we obtain that
0>
b
a
⎛ ⎜
⎝p1t−p2txΔt2q2t−q1tx2σt
⎛ ⎝
yt
p2tyσt
p2tyΔt yt −
p2tyσt yt xΔt
⎞ ⎠
2⎞
⎟ ⎠Δt
>0,
3.14
which is a contradiction, too. Hence, in Case 2, yt has at least one generalized zero on
Case 3. Supposeaσa/2 is a node ofxt, xa >0, xσa< 0,andbis a generalized zero ofxt.Similar to the discussion of3.12, we have
xa
ya
p1axΔaya−p2ayΔaxa
xyaaμ1ap1axσaya−p2ayσaxa p2a−p1axaya
<0,
3.15
which implies
p1axΔaya−p2ayΔaxa
<0. 3.16
iIfbσb/2 is a node ofxt,thenxb<0, xσb>0.Hence, we have3.12, that is,
xb
yb
p1bxΔbyb−p2byΔbxb
<0. 3.17
iiIfbis a zero ofxt,then
xb
yb
p1bxΔbyb−p2byΔbxb
0. 3.18
It follows from3.4andLemma 2.3that
xt
yt
p1txΔtyt−p2tyΔtxt
3.19
is right-increasing ona, b ∩T.Hence, fromiandiithat
xa
ya
p1axΔaya−p2ayΔaxa
< xσa yσa
p1σaxΔσayσa−p2σayΔσaxσa
<0,
3.20
which implies
From3.16,3.21, and2.4, we have
p1xΔy−p2yΔx
Δ
a 1
μa
p1xΔy−p2yΔx
σa−p1xΔy−p2yΔx
a>0. 3.22
Further, it follows from1.1,1.2, product ruleLemma 2.1ii, and3.22that
p1xΔy−p2yΔxΔa q2a−q1axσayσa p1a−p2axΔayΔa>0.
3.23
Ifp1a p2aand fromq2a≥q1a,xσa<0, andyσa>0 we have
q2a−q1a
xσayσa<0. 3.24
This contradicts3.22. Note thatxΔa 1/μaxσa−xa. It follows fromp1a > p2a>0,3.23, and3.24that
yΔa<0. 3.25
On the other hand, it follows fromxtandytare solutions of1.1and1.2that
yσap1axΔa
Δ
q1axσa
0,
xσap2ayΔaΔq2ayσa0.
3.26
Combining the above two equations we obtain
p1axΔa
Δ
yσa−p2ayΔa
Δ
xσa
q1a−q2a
xσayσa 0.
It follows from3.27and2.4that
1
μa
p1σaxΔσa−p1axΔayσa−p2σayΔσa−p2ayΔaxσa
q1a−q2axσayσa
1
μa
p2ayΔaxσa−p1axΔayσa
μ1ap1σaxΔσayσa−p2σayΔσaxσa
q1a−q2axσayσa 0.
3.28
Hence, fromq2a≥q1a, xσa<0, yσa>0,and3.21, we get
p2ayΔaxσa−p1axΔayσa<0. 3.29
By referring toxΔa<0 andp1a> p2a>0,it follows that
yΔa>0, 3.30
which contradictsyΔa<0.
It follows from the above discussion thatyt has at least one generalized zero on
a, b ∩T.This completes the proof.
Remark 3.4. If p1t ≡ p2t ≡ 1, then Theorem 3.3 reduces to classical Sturm comparison
theorem.
Remark 3.5. In the continuous case: μt ≡ 0. This result is the same as Sturm-Picone comparison theorem of second-order differential equationsseeSection 1.
Remark 3.6. In the discrete case:μt≡1. This result is the same as Sturm comparison theorem of second-order difference equationssee8, Chapter 8 .
Example 3.7. Consider the following three specific cases:
0,1 ∩T
0,1 2 ∪
2 3,1 ,
0,1 ∩T
0,1 2 ∪
!
1 2N−1,
1
N−1, 3
2N−1, . . . ,1
"
, N >2,
0,1 ∩Tqk|k≥0, k∈Z∪ {0}, where 0< q <1.
ByTheorem 3.3, we have ifxtandytare the nontrivial solutions of1.1and1.2,a, bare two consecutive generalized zeros ofxt,andp1t≥p2t>0, q2t≥q1t, t∈a, b ∩T,
then yt has at least one generalized zero ona, b ∩T.Obviously, the above three cases are not continuous and not discrete. So the existing results for the differential and difference equations are not available now.
By Remarks3.4–3.6andExample 3.7, the Sturm comparison theorem on time scales not only unifies the results in both the continuous and the discrete cases but also contains more complicated time scales.
Acknowledgments
Many thanks to Alberto Cabada the editor and the anonymous reviewers for helpful comments and suggestions. This research was supported by the Natural Scientific Foundation of Shandong ProvinceGrant Y2007A27,Grant Y2008A28, the Fund of Doctoral Program Research of University of Jinan B0621, and the Natural Science Fund Project of Jinan UniversityXKY0704.
References
1 M. Picone, “Sui valori eccezionali di un parametro da cui dipend un`equazione differenziale linear ordinaria del second ordine,”JMPA, vol. 11, pp. 1–141, 1909.
2 E. Kamke, “A new proof of Sturm’s comparison theorems,”The American Mathematical Monthly, vol. 46, pp. 417–421, 1939.
3 W. Leighton, “Comparison theorems for linear differential equations of second order,”Proceedings of the American Mathematical Society, vol. 13, pp. 603–610, 1962.
4 W. Leighton, “Some elementary Sturm theory,”Journal of Differential Equations, vol. 4, pp. 187–193, 1968.
5 W. T. Reid, “A comparison theorem for self-adjoint differential equations of second order,”Annals of Mathematics, vol. 65, pp. 197–202, 1957.
6 R. Zhuang, “Sturm comparison theorem of solution for second order nonlinear differential equations,”Annals of Differential Equations, vol. 19, no. 3, pp. 480–486, 2003.
7 R.-K. Zhuang and H.-W. Wu, “Sturm comparison theorem of solution for second order nonlinear differential equations,”Applied Mathematics and Computation, vol. 162, no. 3, pp. 1227–1235, 2005. 8 W. G. Kelley and A. C. Peterson, Difference Equations: An Introduction with Applications,
Har-court/Academic Press, San Diego, Calif, USA, 2nd edition, 2001.
9 B. G. Zhang, “Sturm comparison theorem of difference equations,” Applied Mathematics and Computation, vol. 72, no. 2-3, pp. 277–284, 1995.
10 M. Bohner and A. Peterso,Advances in Dynamic Equations on Time Scales, Birkh¨auser, Boston, Mass, USA, 2003.
11 M. Bohner and A. Peterson,Dynamic Equations on Time Scales: An Introduction with Applications, Birkh¨auser, Boston, Mass, USA, 2001.
12 V. Lakshmikantham, S. Sivasundaram, and B. Kaymakcalan,Dynamic Systems on Measure Chains, vol. 370 ofMathematics and Its Applications, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1996.
13 R. P. Agarwal and M. Bohner, “Basic calculus on time scales and some of its applications,”Results in Mathematics, vol. 35, no. 1-2, pp. 3–22, 1999.