Internat. J. Math. & Math. Sci. 35 (1987) 35-45 Vol. i0 No. OSCILLATION THEOREMS FOR SECOND ORDER NONLINEAR DIFFERENTIAL EQUATIONS WITH DEVIATING ARGUMENTS S.R. GRACE Department of Mathematics P. O. Box 1682 University of Petroleum & Minerals Dhahran, Saudl Arabia and B.S. LALLI Devartment of Mathematics University of Saskatchewan Saskatoon, Saskatchewan S7N OWO Canada (Received November 28, 1984 and in revised form March 28, 1986) ABSTRACT, New oscillation criteria for the oscillatory behaviour of the differential (a(t) x(t)) + p(t)x(t) + q(t) f(x[g(t)]) (" 0 d -t) and (a (t) (x (t)) (t)) + p(t)x(t) + q(t) f (x [g (t) ]) 0 are established KEY WORDS AND PHRASES. Oscillation theorems, differential equations, non-oscillatlons. 1980 AMS SUBJECT CLASSIFICATION CODE 34CI0. I. INTRODUCTION. This paper is concerned with the oscillatory behavior of solutions of second order nonlinear damped differential equations with deviating argument of the form (a(t)x(t)) + p(t)x(t) + q(t) f(x[q(t)]) 0 (I.I) and (a (t) ,(x (t) x (t) where a, g, p, q: q(t) not lim g(t) t-> on . [tO, identically -) + p(t)x(t) + q(t) f(x[g(t)]) -> zero , f: R-> R (-", ") are continuous, a(t) > 0, and any ray of the form It*, ) for some t* _> t [0, ), on (1.2) 0 o exist We restrict our attention to those solutions of equations(.T_l)and(l.)which of neighborhood and which are nontrivial in any [t -), t _> t some ray o S. R. GRACE AND B. S. LALLI 36 Such a solution is called oscillatory if it has arbitrarily large zeros, infinity. An equation is called oscillatory if all its otherwise it is called nonoscillatory. solutions are oscillatory. In the study of the second order sublinear differential equation q(t)Ix(t)lasnx(t) x (t) + where q: [tO, -> ) R 0 < ,< i 0, is continuous, (1.3) there are many criteria for oscillation which involve the behavior of the integral of In particular, Belohorec [I] has shown q. that the condition /=sSq(s) ds for some 8 is sufficient for the oscillation of equation [0, ] (1.3), and Kamenev [6] has established that equation (3) is oscillatory if t i lira sup [ s)q(s)ds (t t-> to Recently, Kura [7] has presented a new criterion for the oscillation of equation (1.3) which improves upon those of Belohorec and Kamenev. Kura proved that a sufficient condition for the oscillation of equation (3) is that lira sup t-> t i s) (t sq(s)ds for some S [0,] to These results have been further extended by Philos [8] to a more general equation x (t) + p(t) x(t) + q(t) f(x(t)) p, q: [t where and f O (R)) -> R, f: R-> R is strongly sublinear i.e. 0 0 are continuous, xf(x) > 0, f’(x) > 0 for x du The above results can be applied f+/-0 f-< -. only to ordinary differential equations. The purpose of this paper is to establish some new oscillation criteria for the differential function f equations (I.I) and (1.2). xf(x) > 0 other that In fact, we impose no conditions on the for x Thus our 0 and nondecreasing. results can be applied to superlinear, linear and sublinear differential equations. We also like to mention that we do not stipulate that the function (I.I) and (1.2) is either retarded or advanced. retarded, advanced and equations of mixed-type. 2. THE EUATION assmae that g in equations Hence our theorems hold for ordinary, SECOND ORDER NONLINEAR DIFFERENTIAL EQUATIONS xf(x) > 0 f’(x) > 0 and 37 x # 0 for (2.1) s a- T>_ t o exp( ]a (u) du) ds T T (2.2) Suppose further that there is a dlfferentiable function =)-> (0, (R)) [t o o’: such that o(t) > 0 o(t) < g(t) for t Let conditions (2.1) THEOREM 1. >__ t O and (2.3) limo(t) (2.3) hold and asste that there exists a twice differentiable function p: -> (0, (R)) [t0,) such that - 0 (t) < 0 If lira sup t-> 1 p (t) and 0(t) t f t (t-s) + a (t) < p (t) h a--/ n o(s) q(s)ds < p(t) a(t) for for some (2.4) t > t O n >_ (2.5) i o then equation (l. I) is oscillatory. PROOF. for t _> t I. that Let x(t) be a nonoscillatory solution of equation (I.I), say x(t) > 0 such By a Lemma in [5] and condition (2.2), there exists a t 2 _> t x(t) > 0 x[g(t)] > 0 and for all t (2.6) >__ t 2 Now, define t (t) 0(t) [ t2 x(s) ds f(x [d(s)]) t>_t 2 Then it is easy to verify that ] {(x[)]) t f(x [(t)]) \a(t)/ f[(t)]) 2 + a (a(t) x(t) f"(x[ o(t) ]) (t)o (t) x (t) X[o (t) f’ (x[o(t) 2(x o(t) (2.7) S.R. GRACE and B. 38 t $. LALLI 2 oCt)Ct) x(t) x[o(tl]f’ (x[oCt)]) Using conditions (2.1), (2.3) and (2.6) ue obtain t>_t 2 for f(x[ o(t) ]) <_ and by conditions (2.1), (2.3), (2.4) and (2.6) we have for = >_ t 2 Thus, for ever t >__ t 2 (2.8) we have t (s) ds 2 (t-t2)n= (t2) t + nCt-t 2) n-I (t-s) n-2 u(a) de t2 <_ (t-t2)n .(t2) + n(t-t2)n’l(t 2) eli2. tn t 2 the other hand, for 0 - t > t2 O (-sln e have qlslds 2 a(s) q(s)d= i- 0 SECOND ORDER NONLINEAR DIFFERENTIAL EQUATIONS - and consequently 1 itto -5 S) (t n p(s) a- q(s) ds tit2 <I- )n / ds, a(s) q(s) 0 n’"(t2) n <I t2) < (t 2) + for t >__ t < 39 t2’-I i ---u/ 2 This gives, t lira sup 1 t-> s) (t to which contradicts (2.5). n 0(s) a(s) q(s)ds t2 _< (t 2) + t 0(s) q(s)ds a o This completes the proof of the theorem. For illustration we consider the following examples. EXAMPLE I. x (t) + The differential equation (- l)x(t) t(in[t+sint]) < 1 0 < x(t) has a nonoscillatory solution negative for > t EXAMPLE e, 2. 1 ([ x(t)) Ix[t+sint] 0 sgnx[t+sint] Only the damping coefficient + 2 x(t) 2 4 I +----_Ix[t + cost] a sgnx[t + cost] >_t0 >_ t and 0 (2.10) 2. are satisfied and and 1 x (t) + x(t) has a t3a -3 Ix[t3]l solution nonoscillatory satisfied for p(t) q(t) is negative for x (t) i + is Consider the differential equation (t)- t- I. The conditions of Theorem so, all solutions of equation (2.10) are oscillatory. EXAMPLE 3. The differential equation p(t)- P(t) violating the assumptions of Theorem I. > 0 We let (2.9) t > e and In t. a x(t) + t and t t (t) 0 x(t) All t . 3 except the and t conditions (2.5), condition 1 >_ t o of (2.11) Theorem since are the function On the other hand, the differential equation o t3a-3 Ix[t3] a sgnx[ t 3 is oscillatory by Theorem >_i sgnx[t 3] for p(t) t I/2 0 ,a and i >_ (t) and t t >__ t o 1 (2.12) 3. [9] can describe the One can check that none of the oscillation criteria of If] character of either equation (2.10) for a > 0 or equation (2,12) for oscillatory The following theorem is concerned with the case when condition (2.4) fails. We assume that S.R. GRACE and B. S. LALLI 40 It will be du f(u) ] and (2.13) convenient to make use of the following notation: for any t t we let (t) + a(t) (t, c) 0 c where C2[[t o THEOREM 2. t > t Let >_ 0 (t) 0 ca(t) (0,, )]., ) (2.1) conditions and the function O a(t) O p be as in Theorem >_ ]’(t, i) (2.3) and (2.13) hold and let 0 7 (t, !) o(t) > t for such that <_ 0 t for >_ t o (2.14) and t lira sup < If 0(t) t-> a(u) q(u)duds to tO (2.15) then equation (I. I) is oscillatory. PROOF. for t _> o t fact that Let x(t) be a nonoscillatory solution of equation (1.1), say x(t) > 0 As in the proof of Theorem we get (2.7). Now, using (2.6) and the O(t) ct) < >_ for t t >_ a(t) q(t) +. t we obtain 2 t "it x(t) ds + Y(t, i) t fcx( 2 fcx By condition (2.13), we have C" (t) <- pt) q(t) + a(t) where C Ct) du {(t2)f(U < (t 2) - t a(s) 2 q(s)ds + By the Bonnet theorem, for any t ft y(s, I) 2 l) x(t) f(x(t)) Integrating the above inequality from ). t " (t) + (t xCs) f(x(s)) ds t C0(t 2) + there exists a { vCt 2, I) 2 t C’0(t) > t2, t t 2 f(x(s)) ds y(s i) ; [t2, to t we get x (s) ds f(x(s)) t] such that SECOND ORDER NONLINEAR DIFFERENTIAL EQUATIONS 41 < Thus, for every t >’t (t 2 i) Y(t2, i) du f(u) x(t 2) du f(u) / x(t 2) 2 t (t) < K q(s) ds + C’(t) t2 where (2.16) (t2) C(t2 K + Integrating (2.16) from t a s ft ft 2 2 t q(u)duds y(t2 to 2 I) x(t t f(u" 2 we have _< C-(t) + Kt -(t) + ((t2) C0(t2) , Dividing by O(t) and taking limit superior of both side as t -> we obtain a contradiction to (2.15). This completes the proof of the theorem. It is easy to check that Theorem 2 is not applicable to equation (2.12) if u I. On the other hand, Theorem 2 can be applied in some cases in which Theorem is not applicable. Such a case is described in Example 4 below. EXAMPLE 4. Consider the differential equation IREMARK. x (t) + where c > 0, a x(t) + --ix[g,t)] t > and o(t) lasgnx[g(t)] g(t) with t The conditions of Theorem 2 are satisfied for 0 t o(t) p(t) 0 t L to for > 0, t (2.17) t o and hence equation (2.17) is oscillatory. 3. TRE EQUATION. In order to obtain results for equation (1.2) similar to those in section 2 we assume 0 < C _< (x) <_ C1 S T a exp( T THEOREM 3. in Theorem - x (3.1) p [u) du) ds ca(u) for all T > t o Let conditions (2.1), (2.3), (3.1) and (3.2) hold and let such that (t) < 0 for all and (t) + a_)/(_h (t)\a.()/ < 1 c p__) a(t) for all t > t 0 If condition (2.5). holds, then equation (1.2) is oscillatory. (3.2) p be as (3.3) 42 S.R. GRACE and B. S. LALLI > 0 x(t) Let PROOF x(t) for be a nonoscillatory solution of equation (I2). >_ t o t There exists a t _> t O > x[o(t)] so that 0 The hypotheses of Lemma in [5] are satisfied and hence there exists a Assume that for t 2 _> t t _> t I. such that >_ for all t > 0 x[ (t) and x (t) > 0 t 2 Now, we define Then for every m t t t (t) + a(t) @(x (s)) (s) ds -f(x[o(S)]) (t) a(t) + >_ t 2 we obtain 2 (t) q(t) f(xg(t)]) + f(x[o(t)]) (t) t for ds f(x[)-[i t2 t2 1 a(t) a(t) (x (t)) (t) f(x[o(t)]) (x (t) o(t)(t) *lx(t))f’(x[o(tI])x(t)x[o(t)] f2(x[ o(t) It is easy to verify that (t) <-a(t) q(t) + <--- q(t) a(t) (t) + t >__ a(t)a(t t cla(t f(x[o(t)]) 2 The rest of the proof is similar to that of Theorem and hence is omitted. Next, we present an interesting result, where condition on we replace condition (3.1) by the following one. %(x) > C > 0 for all is weakened, i.e., x (3.4) The result is an immediate consequence of Theorem 3, so we omit the proof. THEOREM 4. Let conditions (2oi), (2.3), (3.2) and (3.4) hold and assume that there exists a function 0 [t0, "), (0, ’)] such that (t) <_ 0 and < (t) + a(t) Pa(t) ) < 0 for t > to (3.5) If condition (2.5) holds, then equation (1.2) is oscillatory. The following examples are illustrative. EXAMPLE 5. Consider the differential equation (i ) x 0 for t >_ t O e (3.6) SECOND ORDER NONLINEAR DIFFERENTIAL EQUATIONS (2.1), (2.3), (2.5) and The conditions The upper bound Equation (3.6) has a nonoscillatory solution (3.1) and (3.3) fail. EXAMPLE 6. > 8 o(t) 8t x(t) p(t) t. conditions Int. Consider the differential equation + where (3.2), of Theorem 3 are satisfied for elXl is undefined and hence both of the function c 43 > 81 -82 0, 0 H2 82 > 0, and a > --+ 2 sgnx[,t sint] and 0 t > to 0 sint] I. We (3.7) J p(t) take All the conditions of Theorem 4 are satisfied and so, and every solution of equation (3.7) is oscillatory. [9] We note that the results in [I] cannot be applied to equation (3.7) since, conditions of the form some of the f--) > K (x) > 0 o=+ lu< (u) x / 0 for (u) ,--=..u< or +0 required in these papers, are not satisfied. EXAMPLE 7. sinx)) ((2 a where (1) g (ii) We > 0 + + g(t) and t-7/61x[g(t)]imsgnx[g(t)] 81 8t o(t) let t +/- 82 g(t) 0 t >_ t O (3.8) 1 satisfies either (i) or (ii): is a nondecreasing continuous functlm for g(t) P(t) Consider the differential equation cost, 8 in case > 0, 81 > (i) and 0 and (t) t 8281 > to with lim g(t) =-. t- _>0. B2 8t in case (ii) and take I/6. The conditions of Theorem 3 are satisfied and so, every solution of equation (3.8) is oscillatory. It is easy to check that Theorem 4 is not applicable to equation (3.8) because condition (3.5) is violated. Next, we consider the differential equation ((I + where a > x2)) 0 Theorem 4 for and p(t) i i Ix[g(t)] +[ x + E g(t) i sgnx[g(t) is as in equation (3.8). 0 t >_ t O 1 (3.9) Equation (3.9) is oscillatory by 1. It is easy to verify that Theorem 3 fails to apply to equation (3.9), since condition (3.1) is not satisfied. REMARK. The above examples illustrate that our results apply to superlinear, linear or sublinear damped differential equations. Moreover, since we impose no restrictions applicable on the function g in equations (I.I) and (1.2), our results are to ordinary, retarded, advanced and equations of mixed type. S.R. GRACE and B. S. LALLI 44 We that believe the of behavior oscillatory equation (3.9) (3.7) is not deducible from any other known oscillation criteria. Finally, we give the following oscillation criterion which is similar to the one Here we omit the proof. in Theorem 2. Let THEOREM 5. t t conditions (2.1), (2.3), (3.1) and (3.2) o(t) hold, > for t and O du f(u) and < f(u) -((t, >_ 0 (t) I) >__ C C2[[t0,(R)), p Assume that there exists a function cu < 7Ct, c-I) < 0 0 (3. i0) (0,(R))] such that t > t for o (3.11) and t lira sup t-> < condition (2.15) holds, than equation (1.2) is oscillatory. The following example is illustrative. EXAMPLE 8. Consider the differential equation i +1 x +--IxCgCt)]lsgnx[g(t)] sinx)) ((2 0 t t > t and It O P(t) I. > a where is any nondecreasing continuous function with check to that g(t) > t for the conditions of Theorem 5 are satisfied with and hence equations (3.12) is oscillatory. t p(t) If g(t) easy is t > tO > 0 0, (2.2) then conditions and (3.2) take the form fas) ds and condition (3.2) can be replaced by condition (3.4). 2. (I.I) and (1.2) In that case, we can introduce a continuous, The results of this paper can be applied to equations of the form when f is not a monotonic function. nondecreasing function xF(x) > 0 ar.] F f() on R such that for C >_ x # 0 (3.13) For illustration we can consider the following differential equation. (t.Cx)) where g + t f(x[gCt)]) is as in Example 7, 0 f t >_ t o > 0 1 0 < e < (3.14) is any continuous, nondecreaslng function on R SECOND ORDER NONLINEAR DIFFERENTIAL II EQUATIONS 45 => xf(x)> 0 for x # 0 e.g. f(x) sgnx, 0 or f(x) sinhx or f is any continuous function on R satisfying condition (3.13) e.g. sinx x e f(x) sgnx, e 0...etc. and is any continuous function on with > satisfying condition (3.4) e.g. If take we /, p(t) the + x 2 or e x or In(e + x 2) (x) or 2 +/- R sinx. of Theorem 4 are satisfied and thus all conditions solutions of equation (3.14) are oscillatory. In the case (x) I, oscillatory solution x(t) 3. e --0, f(x) x and g(t) t, equation (3.14) has the sinlnt. The results of this paper are presented in a form which is essentially new. REFERENCES I. BELOHOREC, S., Two Remarks on the Equation. Acta. Fac. Differential Properties of Re num. Natur. Solutions of a Comenian. Univ. Nonlinear Math. 22 (1969), 19-26. 2. GRACE, S. R. and LALLI, B. S., Functional Differential Oscillation Theorems for Nonlinear Second Order Equations with Damping. Bull. Inst. Math. Acad. Sinica, 13 (1985), 192-193. 3. GRACE, S. R. and LALLI, B. S., Oscillation Theorems for Differential Equations. J. Math. Anal. Appl. (to appear). 4. GRACE, S. R. and LALLI, B. S., Oscillation of Solutions of Damped Nonlinear Second Order 5. Functional Differential Equations. Second Order Nonlinear Bull. Inst. Math. Acad. Sinica, 12 (1984), 5-9. GRACE, S.R., LALLI, B.S. and YEH, C.C., Oscillation Theorems for Nonlinear Second Order Differential Equations With a Nonlinear Damping Term. SlAM J. Math. Anal., 15 (1984), 1082-1093. 6. KAMENEV, I. V., Some I0 (1971), 129-134. 7. KURA, T., Oscillation Equation. Specifically Nonlinear Oscillation Theorems. Mat. Zametki., Theorems For a Second Order Sublinear Ordinary Differential Proc. Amer. Math. Soc., 84 (1982), 535-538. 8. PHILOS, Ch. G., Oscillation of Sublinear Differential Equations of Second Order. Nonlinear Anal. Theory Method and Appl., 7 (1983), 1071-1080. 9. WONG, J. S. W., On Second Order Nonlinear Oscillation, Funkcial Ekvac., II (1968), 207-234.