Interval criteria for oscillation of second-order differential equations with mixed nonlinearities

https://doi.org/10.1016/j.amc.2007.08.042Get rights and content

Abstract

In this paper, we establish interval criteria for oscillation of the second order forced ordinary differential equations with mixed nonlinearities:(p(t)x)+q(t)x+i=1nqi(t)|x|αisgnx=e(t),where p(t), q(t), qi(t), e(t) are continuous functions defined on [0,∞), p(t) is positive and differentiable, α1 >  > αm > 1 > αm+1 >  > αn > 0(n > m  1). No restriction is imposed on the potentials q(t), qi(t) and e(t) to be nonnegative.

Introduction

In this paper, we investigate the oscillatory behavior of solutions of the second order nonlinear differential equation(p(t)x)+f(t,x)=0,t0,where p(t) is positive and differentiable and f(t,x) is of the form:f(t,x)=q(t)x+i=1nqi(t)|x|αisgnx,where q(t), qi(t)  C[0,∞), and α1 >  > αm > 1 > αm+1 >  > αn > 0.

As usual, a solution x(t) of (1) is said to be oscillatory if it is defined on some ray [T,∞) with T  0 and has unbounded set of zeros. Eq. (1) is said to be oscillatory if all solutions extendable throughout [0,∞) are oscillatory. When f(t, x) takes the form (2), it is known that if q(t), qi(t) are also positive and continuously differentiable then all solutions of (1) are extendable throughout [0,∞). However, when q(t), qi(t) change signs as t tends to infinity, it is also known that Eq. (1) can have solutions with finite escape time, i.e. it can become infinite at some finite t. See Coffman and Wong [6]. We shall for simplicity confine our discussion only to extendable solutions throughout this paper.

For the particular case when p(t)  1, q(t)  0, and n = 1 in (2), Eq. (1) reduces to the familiar Emden–Fowler equationx+q1(t)|x|α1sgnx=0.When α1 > 1, Eq. (3) is known as the superlinear equation and when 0 < α1 < 1, it is known as the sublinear equation. The oscillation of Eq. (3) has been the subject of much attention during the last 50 years, see the seminal book by Agarwal, Grace and O’Regan [2]. Most of the results on oscillation of (3) are valid for the more general equationx+q(t)f(x)=0,where xf(x) > 0 for x  0 and f(x) satisfies certain conditions of superlinearity and sublinearity (see [2]). In particular, f(x) can be finite sum of powers of x and if there exist in this sum exponents of x which are both greater than and less than 1, then Eq. (4) is known as Emden–Fowler equation of the mixed type. When f(t,x) in (1) takes the form of (2) and Eq. (1) is of a mixed type, results on oscillation are more or less the same when q(t), qi(t), i = 1, 2 …, n, are non-negative. This is however not the case when q(t), qi(t) are oscillatory. When n = 1, it is known (see Butler [4]) that Eq. (3) is oscillatory if q1(t) is periodic and of mean value zero, e.g. sint or cost. When f(x) is a finite sum of powers of x, the more general equation is also oscillatory as shown in a subsequent paper by Butler [5] (see also [12]). However, for Eq. (1) with f(t, x) as given in the form of (2) there seems to us no known oscillation criterion applicable to the simple equation of the mixed type:x+sint|x|α1sgnx+cost|x|α2sgnx=0,where α1 > 1 and 0 < α2 < 1.

On the other hand, if we consider the forced equation(p(t)x)+q(t)x+i=1nqi(t)|x|αisgnx=e(t),where e(t) is itself an oscillatory function, we also wish to establish oscillation theorems when n > 1. When n = 1, this approach was initiated by Agarwal and Grace [1, pp. 244–249] for higher order equations and subsequently developed in papers of Ou and Wong [14], Sun and Agarwal [22], [23], Sun and Saker [24], Yang [29], and Yang [30].

For the simple case when n = 1, p(t)  1, q(t)  0 and q1(t) is non-negative in (5), one can use a technique first introduced by Kartsatos [8], [9] by assuming that e(t) is the second derivative of an oscillatory function to obtain oscillation criteria for Eq. (5) as described in the earlier work [27]. See also previous results in Keener [10], Rankin [16], Skidmore and Leighton [17], Skidmore and Bowers [18] and Teufel [26] for special cases of the forced Eq. (5).

When the assumption that e(t) is the second derivative of an oscillatory function is not imposed, one can develop interval oscillation criteria for special cases of (5), in case qi(t)  0 for all i = 1, 2, …, n, see El-Sayed [7], Sun, Ou and Wong [21] and Wong [28], in case q(t)  0 and n = 1 see Nazr [13] when α1 > 1 and Sun [19], Sun and Wong [20] when 0 < α1 < 1.

In this paper, we will employ the method in Kong [11] and the arithmetic-geometric mean inequality (see [3]) to establish several interval oscillation criteria for the unforced Eq. (1) and the forced Eq. (5), respectively. In the former case our results are generalizations of the main results in [11]. Examples are given to illustrate our results when compared with known results on Emden–Fowler equations. We hope to kindle reader’s interest in further research on the oscillation of equations of mixed type, which arise for example in the growth of bacteria population with competitive species.

Section snippets

Main results

We will need the following lemmas that have been proved in our recent paper [25]:

Lemma 1

Let {αi}, i = 1, 2, , n, be the n-tuple satisfying α1 >  > αm > 1 > αm+1 >  αn > 0. Then there exists an n-tuple (η1, η2, , ηn) satisfyingi=1nαiηi=1,which also satisfies eitheri=1nηi<1,0<ηi<1ori=1nηi=1,0<ηi<1.

Lemma 2

Let u, A, B, C and D be positive real numbers. Then

  • (i)

    Auγ + B  γ(γ  1)1/γ1 A1/γB11/γu, γ > 1;

  • (ii)

    Cu  Duγ    1)γγ/(1γ) Cγ/(γ1)D1/(1γ), 0 < γ < 1.

Remark 1

For a given set or exponents {αi} satisfying α1 >  > αm > 1 > αm+1 >  αn > 0, Lemma 1 ensures the

Examples

In this section, we will work out two numerical examples to illustrate our main results.

Example 1

Consider the following equation:x(t)+ksint|x|α1sgnx+lcost|x|α2sgnx=-mcos2t,t0,where k, l, m are positive constants, α1 > 1 and 0 < α2 < 1. According to the direct computation, we have Q(t)=k0cos2tη0(sint)η1(cost)η2, where k0=(η0/m)-η0(η1/k)-η1(η2/l)-η2, η0 can be any positive number satisfying 0 < η0 < (α1  1)/α1, η1 and η2 satisfy condition (6). For any T  0, we can choose a1 = 2, b1 = a2 = 2 + π/4, b2 = 2 + π/2, c1 = 2 + π

Acknowledgements

This work was supported by the National Natural Science Foundations of China (10771118 and 10671105) and the Natural Science Foundations of Shandong Province of China (Q2006A01 and Y2005A06).

References (30)

  • Q. Yang

    Interval oscillation criteria for a forced second order nonlinear ordinary differential equations with oscillatory potential

    Appl. Math. Comput.

    (2003)
  • X. Yang

    Forced oscillation of nth order nonlinear differential equations

    Appl. Math. Comput.

    (2003)
  • R.P. Agarwal et al.

    Oscillation Theory for Difference and Functional Differential Equations

    (2002)
  • R.P. Agarwal et al.

    Oscillation Theory for Second Order Linear, Half-linear, Superlinear and Sublinear Dynamic Equations

    (2002)
  • E.F. Beckenbach et al.

    Inequalities

    (1961)
  • Cited by (34)

    • Interval criteria for forced oscillation with nonlinearities given by RiemannStieltjes integrals

      2011, Computers and Mathematics with Applications
      Citation Excerpt :

      Clearly, Theorems 2.3 and 2.4 not only cover but also unify and improve Theorems 1 and 2 of [2,16].

    View all citing articles on Scopus
    View full text