A LIMIT-POINT CRITERION FOR A SECOND-ORDER LINEAR DIFFERENTIAL OPERATOR IAN KNOWLES

Similar documents
Contraction Mapping Principle Approach to Differential Equations

Mathematics 805 Final Examination Answers

REAL ANALYSIS I HOMEWORK 3. Chapter 1

EXISTENCE AND UNIQUENESS OF SOLUTIONS FOR A SECOND-ORDER ITERATIVE BOUNDARY-VALUE PROBLEM

Convergence of Singular Integral Operators in Weighted Lebesgue Spaces

5.1-The Initial-Value Problems For Ordinary Differential Equations

Green s Functions and Comparison Theorems for Differential Equations on Measure Chains

INTEGRALS. Exercise 1. Let f : [a, b] R be bounded, and let P and Q be partitions of [a, b]. Prove that if P Q then U(P ) U(Q) and L(P ) L(Q).

1.0 Electrical Systems

Application on Inner Product Space with. Fixed Point Theorem in Probabilistic

e t dt e t dt = lim e t dt T (1 e T ) = 1

GENERALIZATION OF SOME INEQUALITIES VIA RIEMANN-LIOUVILLE FRACTIONAL CALCULUS

Motion. Part 2: Constant Acceleration. Acceleration. October Lab Physics. Ms. Levine 1. Acceleration. Acceleration. Units for Acceleration.

Hermite-Hadamard-Fejér type inequalities for convex functions via fractional integrals

ENGR 1990 Engineering Mathematics The Integral of a Function as a Function

f t f a f x dx By Lin McMullin f x dx= f b f a. 2

3. Renewal Limit Theorems

1. Introduction. 1 b b

September 20 Homework Solutions

Journal of Mathematical Analysis and Applications. Two normality criteria and the converse of the Bloch principle

ON NEW INEQUALITIES OF SIMPSON S TYPE FOR FUNCTIONS WHOSE SECOND DERIVATIVES ABSOLUTE VALUES ARE CONVEX

ASYMPTOTIC BEHAVIOR OF INTERMEDIATE SOLUTIONS OF FOURTH-ORDER NONLINEAR DIFFERENTIAL EQUATIONS WITH REGULARLY VARYING COEFFICIENTS

MATH 124 AND 125 FINAL EXAM REVIEW PACKET (Revised spring 2008)

Some Inequalities variations on a common theme Lecture I, UL 2007

How to Prove the Riemann Hypothesis Author: Fayez Fok Al Adeh.

FURTHER GENERALIZATIONS. QI Feng. The value of the integral of f(x) over [a; b] can be estimated in a variety ofways. b a. 2(M m)

Positive and negative solutions of a boundary value problem for a

How to prove the Riemann Hypothesis

Integral Transform. Definitions. Function Space. Linear Mapping. Integral Transform

ON NEW INEQUALITIES OF SIMPSON S TYPE FOR FUNCTIONS WHOSE SECOND DERIVATIVES ABSOLUTE VALUES ARE CONVEX.

4.8 Improper Integrals

An integral having either an infinite limit of integration or an unbounded integrand is called improper. Here are two examples.

0 for t < 0 1 for t > 0

Minimum Squared Error

Minimum Squared Error

Solutions of half-linear differential equations in the classes Gamma and Pi

Yan Sun * 1 Introduction

New Inequalities in Fractional Integrals

The Asymptotic Behavior of Nonoscillatory Solutions of Some Nonlinear Dynamic Equations on Time Scales

ON THE OSTROWSKI-GRÜSS TYPE INEQUALITY FOR TWICE DIFFERENTIABLE FUNCTIONS

MTH 146 Class 11 Notes

EXERCISE - 01 CHECK YOUR GRASP

MAT 266 Calculus for Engineers II Notes on Chapter 6 Professor: John Quigg Semester: spring 2017

Procedia Computer Science

Approximation and numerical methods for Volterra and Fredholm integral equations for functions with values in L-spaces

Magnetostatics Bar Magnet. Magnetostatics Oersted s Experiment

ANSWERS TO EVEN NUMBERED EXERCISES IN CHAPTER 2

Example on p. 157

white strictly far ) fnf regular [ with f fcs)8( hs ) as function Preliminary question jointly speaking does not exist! Brownian : APA Lecture 1.

Probability, Estimators, and Stationarity

On Hadamard and Fejér-Hadamard inequalities for Caputo k-fractional derivatives

Chapter Direct Method of Interpolation

Honours Introductory Maths Course 2011 Integration, Differential and Difference Equations

( ) ( ) ( ) ( ) ( ) ( y )

Non-oscillation of perturbed half-linear differential equations with sums of periodic coefficients

CALDERON S REPRODUCING FORMULA FOR DUNKL CONVOLUTION

MODULE 3 FUNCTION OF A RANDOM VARIABLE AND ITS DISTRIBUTION LECTURES PROBABILITY DISTRIBUTION OF A FUNCTION OF A RANDOM VARIABLE

Refinements to Hadamard s Inequality for Log-Convex Functions

AN EIGENVALUE PROBLEM FOR LINEAR HAMILTONIAN DYNAMIC SYSTEMS

Copyright by Tianran Geng 2017

Fractional Calculus. Connor Wiegand. 6 th June 2017

Physics 2A HW #3 Solutions

A Kalman filtering simulation

Average & instantaneous velocity and acceleration Motion with constant acceleration

IX.2 THE FOURIER TRANSFORM

Hardy s inequality in L 2 ([0, 1]) and principal values of Brownian local times

Analytic solution of linear fractional differential equation with Jumarie derivative in term of Mittag-Leffler function

4 Sequences of measurable functions

Applications of Prüfer Transformations in the Theory of Ordinary Differential Equations

Solutions to Problems from Chapter 2

arxiv: v1 [math.ca] 15 Nov 2016

graph of unit step function t

The solution is often represented as a vector: 2xI + 4X2 + 2X3 + 4X4 + 2X5 = 4 2xI + 4X2 + 3X3 + 3X4 + 3X5 = 4. 3xI + 6X2 + 6X3 + 3X4 + 6X5 = 6.

SOLUTIONS TO ASSIGNMENT 2 - MATH 355. with c > 3. m(n c ) < δ. f(t) t. g(x)dx =

15. Vector Valued Functions

arxiv: v1 [math.pr] 24 Sep 2015

Asymptotic relationship between trajectories of nominal and uncertain nonlinear systems on time scales

Generalized Snell envelope and BSDE With Two general Reflecting Barriers

FM Applications of Integration 1.Centroid of Area

An Introduction to Malliavin calculus and its applications

Positive continuous solution of a quadratic integral equation of fractional orders

..,..,.,

PHYSICS 1210 Exam 1 University of Wyoming 14 February points

14. The fundamental theorem of the calculus

BOUNDED VARIATION SOLUTIONS TO STURM-LIOUVILLE PROBLEMS

(b) 10 yr. (b) 13 m. 1.6 m s, m s m s (c) 13.1 s. 32. (a) 20.0 s (b) No, the minimum distance to stop = 1.00 km. 1.

P441 Analytical Mechanics - I. Coupled Oscillators. c Alex R. Dzierba

For the reaction, R P, the is given by,

On the Pseudo-Spectral Method of Solving Linear Ordinary Differential Equations

1 jordan.mcd Eigenvalue-eigenvector approach to solving first order ODEs. -- Jordan normal (canonical) form. Instructor: Nam Sun Wang

Some basic notation and terminology. Deterministic Finite Automata. COMP218: Decision, Computation and Language Note 1

A Structural Approach to the Enforcement of Language and Disjunctive Constraints

Dynamic Systems and Applications 12 (2003) A SECOND-ORDER SELF-ADJOINT EQUATION ON A TIME SCALE

Supplement for Stochastic Convex Optimization: Faster Local Growth Implies Faster Global Convergence

3 Motion with constant acceleration: Linear and projectile motion

Think of the Relationship Between Time and Space Again

on the interval (x + 1) 0! x < ", where x represents feet from the first fence post. How many square feet of fence had to be painted?

An Integral Two Space-Variables Condition for Parabolic Equations

MATH 5720: Gradient Methods Hung Phan, UMass Lowell October 4, 2018

A new model for limit order book dynamics

Transcription:

A LIMIT-POINT CRITERION FOR A SECOND-ORDER LINEAR DIFFERENTIAL OPERATOR j IAN KNOWLES 1. Inroducion Consider he forml differenil operor T defined by el, (1) where he funcion q{) is rel-vlued nd loclly Lebesgue inegrble on he rel inervl /. A soluion y{) of he differenil equion iy{) = 0 is sid o be squreinegrble n end-poin b of / if here is neighbourhood N of b such h /()l < oo. / n N If every soluion of xy{) = 0 is squre-inegrble n end-poin b of/, hen we sy h he operor T is of limi-circle ype b; oherwise x is of limi-poin ype b. In his pper we furher reduce he lrge number ofindependen limi-poin crieri wih resul which simulneously generlizes he well-known crieri of Brinck [1], Hrmn [4] nd Sers'[6]. Noe h he firs of hese resuls ws originlly given s condiion for selfdjoinness. In priculr, Brinck proved h he closed liner operor T in L 2 (- oo, oo) defined by &i(t) = {/el 2 (-oo, oo): T/GL 2 (-OO, oo),/'is loclly bsoluely coninuous} T/=T/ for is self-djoin provided h here exiss sufficienly smooh (see ler) posiive funcion w() wih j w()q()d ^ C (2) for ll inervls c (-oo, oo) wih /() ^ 1, where l() is he lengh of he inervl. Now, since T is self-djoin if nd only if he forml operor T is of limi-poin ype boh oo nd oo, n equivlen semen of Brinck's resul sys h T will be of limi-poin ype oo provided h here exiss smooh posiive funcion w() sisfying (2) for ll inervls <= [, oo) wih 1() < 1, for some > oo. In he ligh of his informion, we shll henceforh resric our enion o he inervl/ = [, oo) where > co. 2. Preliminries LEMMA 2. // w is posiive nd loclly bsoluely coninuous on he hlf-line [, oo) nd sisfies \w'(s)\ ^ 1 lmos everywhere on [, oo), hen here exiss sequence Received 13 April, 1973. [. LONDON MATH. SOC. (2), 8 (1974), 719-727]

720 IAN KNOWLES { k } such h 0 =, k -* oo s k -* oo nd Proof. We firs prove h I {w(s)}~ 1 ds= oo. (4) Assume h he bove inegrl is finie. Then for ll s ^ log [w(s)/w()] = I w'(u)/w(u)du s oo < I {^(M)}" 1^ since \w'\ ^ 1 <. Thus {W()}" 1 is bounded wy from zero nd his conrdics our iniil ssumpion. Consider now he funcion A(0= j MsT'ds. I is cler from (4) nd he inermedie vlue heorem h here exiss poin h > h = Similrly, by considering he funcion / 2 (0= f Ms)} one cn find poin 2 > x wih i / 2 (' 2 )= nd so on. The sricly incresing sequence so produced is obviously unbounded, from he coninuiy of w.

A LIMIT-POINT CRITERION FOR A SECOND-ORDER LINEAR DIFFERENTIAL OPERATOR 721 LEMMA 2. (Gnelius [3]). ///^ 0 nd g re funcions of bounded vriion on he compc inervl K <= R, hen f f(s)dg(s) ^ (inf/+vrf) sup f dg(s) \ K K / CK K where he supremum is ken over ll inervls <= K. The proof of his lemm will be omied. 3. The min resul THEOREM. // here exiss posiive, loclly bsoluely coninuous funcion w(s) defined on [, oo) nd such h () w(s)ds = oo, (b) w' is essenilly bounded on [, oo), (c) here exiss consn C > Ofor which f w(s)q(s)ds ^ C if f M*)}" 1 ds ^ 1, hen he forml operor T defined by equion (1) is of limi-poin ype oo. Proof I will be sufficien o show h no every rel soluion of xy() = 0 is in l}[, oo). We proceed by conrdicion. Assume wihou loss of generliy h \w'(s)\ < 1 lmos everywhere on [, oo) nd le f Gl}[, oo) be he rel soluion sisfying f x {) = 0 nd fi() = 1; le f 2 el3[, oo) be he rel soluion sisfying f 2 () = 1 nd f 2 '{) = 0. Using he sequence { k } given in Lemm 1, we define funcions 0 fc (s), k = 0, 1,..., on [, oo) by 'k+ 1 w' 1 (r)dr, k ^ s < / fc+1, 5 I is esily seen h ech funcion 0 & is loclly bsoluely coninuous, nd sisfies O<0 fc^l, (5) w(s)\^) k '(s)\ < 1.e. on [, oo), (6) (f) k (s) - 1 s k -+ co for ll se [o, oo). (7) 10

722 IAN KNOWLES Le (f) be ny one of hese funcions nd define \//(s) = <f) 2 (s) w 2 (s). Then, 0= /,(#(s)t/,(s)ds 'k+l 'k + 1 = j q(s)hs)a 2 (s)ds- j {f x '{s)} 2 il,(s)ds fer inegring by prs, f k + 1 "2 /i'(5)/( 5 ){^( S )f (5)>V 2 (5) +»V(S)»V'( 5 ) 'k + 1 'k + 1 'k + 1 < q(s)ilj(s)f 2 l (s)ds- j {f l '(s)} 2 Hs)d3 +4 / 1 '(s)/ 1 (s) 0(s)w(s)^s (8) using (5) nd (6) nd condiion (b), where f k+l f l+l f. 2 v f f 2 A l l i=o k / \ C < X I inf w fi 2 <f> 2 + vr w'/i 2^2) sup wq by Lemm 2, * /,. r +1 ^ C 2 I inf wfi <p w ' = 0 \[r (>»,+,] by equion (3) nd condiion (c), (< j 'k+

A LIMIT-POINT CRITERION FOR A SECOND-ORDER LINEAR DIFFERENTIAL OPERATOR 723 Finlly, subsiuing his esime ino he inequliy (8), nd hen using he Cuchy- Schwrz inequliy gives Hs)Ui(s)} 2 ds ^ 0(1)11 + **(s) 1/,'Os)! U which proves h he expression on he lef-hnd side of he ls inequliy mus be bounded by number which is independen of he priculr \j = \/ k h we use. Hence, by Fou's lemm, leing k -> oo Similrly, oo Now i is cler h (w 2 (s){f i '(s)} 2 ds ^ liminf f w 2 {s)(l) k 2 (s){f l '(s)} 2 ds k-»oo <. oo»v 2 (S){/ 2 '(5)} 2 <OO. {wf x ')f 2 -{wf 2 ')h=w nd, since we hve shown h ech of he four fcors on he righ-hnd side is in L 2 [, oo), we clerly hve w(s)ds < oo. This conrdics condiion () nd complees he proof. COROLLARY 1 (Brinck [1]). // here exiss posiive funcion w(s) defined on [, oo) nd sisfying () w, w' nd w" re ll bounded nd coninuously differenible on [, oo) nd oo w(s)ds = oo. (P) here exiss consn C > 0 wih f w(s)q(s)ds ^ C for ll inervls <= [, oo) wih /() ^ 1, hen T 15 of limi-poin ype oo. j

724 IAN KNOWLES Proof. Assume wihou loss of generliy h w < 1. We hve only o prove h condiion (c) of he heorem is sisfied. This, however, follows immediely from condiion (/?) bove nd he inequliy K)< {w(s)y 1 ds. The crierion of Sers [6] is lso specil cse. I is in fc equivlen o he following COROLLARY 2. // here exiss posiive funcion w(s) which is loclly bsoluely coninuous on [, oo) nd sisfies () w' is essenilly bounded on [, oo) nd uu s = oo, (/?) w 2 q is essenilly bounded bove on [, oo), hen T is of limi-poin ype oo. Proof Assume wihou loss of generliy h he upper bound in condiion (/) is 1. Then for ny inervl, f w(s)q(s)ds^ I {^(s)}" 1^ / j nd condiion (c) of he heorem is once gin sisfied. COROLLARY 3 (Hrmn [4]). // here exiss posiive non-decresing funcion u() defined on [, oo) nd rel number 9, 0 < 6 < 1, such h () (w(r)} dr = oo, (f$) here exiss consn B such h for ll e [, oo) u(+0/u()) ^ (y) for ll, se [, oo) such h 0 ^ -s f/?e«t is of limi-poin ype oo. Bu{\ < 0/«(O «(r)dr< i/(0, s Proof. Our objecive here will be o consruc funcion w() which sisfies he requiremens of he previous heorem. Observe firs of ll h if I u(r)dr s% 0, ^ s < / < oo,

A LIMIT-POINT CRITERION FOR A SECOND-ORDER LINEAR DIFFERENTIAL OPERATOR 725 hen u(s) ( s) ^ 0 nd herefore ^ s+6fu(s). Consequenly, by condiion (/?), i.e. w(0 < u(s+9/u(s)) ^ Bu(s), (/?') u()/u(s) < B if s < / nd f w(r)dr ^ 0. Now, since u is non-decresing, we cn immediely sser h u(r)dr = oo. Hence, by he mehod of Lemm 1, here mus exis sequence { k } wih * 0 =, k -+ oo s k -> oo nd 'k+ 1 u(r)fr = 0, fc = O,l,... (9) k Define /i k = / k+1 k for A; = 0, 1,... Then by equion (9) nd condiion (/?'), for A: = 0, 1,... The funcion w(s) is defined on [ k, k+1 ] by (10) The piecewise liner funcion so defined is obviously posiive, loclly bsoluely coninuous nd monoone non-incresing over [, oo). Also, from inequliy (10), nd ^ B l.e. on [, oo) 'k+ 1 f w(s)ds = Z f S A^k H '('k+ I) s^6 vv is decresing, = oo. k+i f {«(s)}- l ds by(js') Our finl sk is o show h w sisfies condiion (c) of he heorem. Observe h if s e [ k, k+1 ] hen w(s) M(S) ^ w( k ) u(s) = 6B~ 2 u(s)/u( k ) ^ 6B' 1. Furhermore, his is rue for every vlue of k; we my herefore conclude h he

726 IAN KNOWLES bove inequliy holds for ll s e [, oo). Now, choose ny inervl [s, ] wih I follows from he bove inequliy h nd hence h u(s) (/ s) ^6B~ l. Rerrnging nd using condiion (/?') we find h s^ 9/u(), from which we infer (vi condiion (y)) h Thus by Lemm 2, jq(r)dr^u(). s ' r * w(r)q(r)dr ^ infw+ \w'(r)\dr sup q(r)dr [s,] Kc[s,] K by (ll), (11) nd he proof is complee. As finl remrk, he uhor feels h similr resul should hold for he generl operor where he funcion p() is posiive nd is reciprocl is loclly Lebesgue inegrble on [, oo); his would hen generlize he resul of Levinson [5] nd possibly even h of Everi [2]. Unforunely, s he mehod used in he heorem does no immediely crry hrough o he generl operor, his will hve o remin conjecure for he presen. The uhor wishes o hnk Professor P. Hrmn for severl helpful suggesions. References 1. 1. Brinck, " Self-djoinness nd specr of Surm-Liouville operors ", Mh. Scnd., 7 (1959), 219-239. 2. W. N. Everi, " On he limi-poin clssificion of second-order differenil operors ", /. London Mh. Soc, 41 (1966), 531-534. 3. T. Gnelius, " Un heoreme uberien pour l rnsformion de Lplce ", C. R. Acd. Sci. Pris, 242 (1956), 719-721. 4. P. Hrmn, " On dichoomies for soluions of n-h order liner differenil equions ", Mh. Ann., 147 (1962), 378-421.

A LIMIT-POINT CRITERION FOR A SECOND-ORDER LINEAR DIFFERENTIAL OPERATOR 727 5. N. Levinson, " Crieri for he limi-poin cse for second-order liner differenil operors ", Csopis pro pisovni memiky fysiky, 74 (1949), 17-20. 6. D. B. Sers, " Noe on he uniqueness of he Green's funcions ssocied wih cerin differenil equions ", Cnd.. Mh., 2 (1950), 314-325. School of Mhemicl Sciences, Flinders Universiy of Souh Ausrli, Bedford Prk, Souh Ausrli. Curren ddress: Deprmen of Mhemics, Universiy of he Wiwersrnd, ohnnesburg, Souh Afric.