Convergence in Nonlinear Systems With A Forcing Term: 1 Preliminars

Revista de Matematica: Teor y Aplicaciones 4(1): 14 (1997) a
convergence in nonlinear systems with a forcing term

J. E. Napoles1 A. I. Ruiz2
Abstract The problem of convergence, as t +, of solutions of the system (1) is considered. It is assumed that the functions , f and g are of class C 1 for all values of their arguments, furthermore g (x) > 0, f (x) r > 0, 0 < n (y) N < + and the functions a(t) and p(t) are continuous on [0, +) with 0 < a a(t) A < + and p(t) 0. Resumen Se considera el problema de la convergencia, cuando t +, de las soluciones del sistema (1). Se asume que las funciones , f y g son de clase C 1 para todos los valores de sus argumentos y que adems g (x) > 0, f (x) r > 0, 0 < n (y) N < + a y las funciones a(t) y p(t) son continuas sobre [0, +) con 0 < a a(t) A < + y p(t) 0.
Preliminars
x y = (y) f (x), = p(t) a(t)g(x), (1)
We consider the system:
where the functions involved are of class C 1 for all values of their arguments, we also assume that g (x) > 0, f (x) r > 0, for all x R and 0 < n (y) N < +, for all y R and the functions a(t) and p(t) are continuous functions on [0, +) with 0 < a a(t) A < + and p(t) 0. In this paper we study the problem of convergence, when t +, of all solutions of the system (1). If the functions involved in this system are such that (y) = y and a(t) 1, the system reduces to the well known Linards equation with forcing term: e x + f (x)x + g(x) = p(t). (2)
The purpose of this note is to generalize, to the system (1), the results of [2] concerning to the convergence of all solutions of (2) to a bounded solution, under the earlier conditions.
Department of Mathematics and Computation, ISP Jos de la Luz y Caballero, Holgu e n 81000, Cuba 2 Department of Mathematics, Universidad de Oriente, Santiago de Cuba 5, CP 90500, Cuba
1
j. e. napoles a. i. ruiz
Results
Let ((t), (t)) be a bounded solution of (1). If we set v(t) = x(t)(t), w(t) = y(t)(t) we obtain the system: v w = (w + ) () [f (v + ) f ()] , = a(t) [g(v + ) g()] . (3)
In the half plane v > 0 we consider lines w + kv = const., where k 0 for nw > av and k < 0 for nw av. We nd conditions under which trajectories of (3) intersect these lines at points where w = lv, in the direction of the origin; this means that w + k is decreasing on the solutions of system (3), i.e. there exist positive numbers P and Q such that, all the solutions of (3) cannot leave the domain |v| < P , |w| < Q by crossing the boundary v = P, w = Q when t +. Dierentiating w + kv with (3) we obtain:
z u
k
z
(r)dr
u
a(t)g (s) kf (s) ds = v a(t)g (u) kf (u) + kl (z) ,
(4)
where = u u u = + v, = z z z = + w. It follows from this that the conditions for intersection in the desired direction takes the form: k< Let: H(l) = inf R(l, t, u, z)
t,u,z
a(t)g (u) := R(l, t, u, z). l (z) f (u)
(5)
we have that the inmum take place when 1 l > r, for u such that f (u) < l1 and 1 = inf z (z). Hence we have the following: Lemma 1 The function R(l, t, u, z) reaches the inmum in the absolute extremes of the function (z). Let U (l) H(l) be any piecewise smooth function satisfying U (l) 0 for 1 l > r, U (l) < 0 for 1 l r and lim Ul(r/N ) U (l) = . Consider the homogeneous dierential equation: dw + U (w/v) = 0. dv (6)
Let the integral curve of equation (6) which passes through the point v = 0, w = w0 > 0 and which lies in the right half-plane intersect the w-axis again at w1 < 0 (if there is no such intersection, take w1 = +). Theorem 1 If w1 < w0 , then (x(t), y(t)) ((t), (t)) as t +, where (x(t), y(t)) is an arbitrary solution of (1).
convergence in nonlinear systems with a forcing term
Proof. Let : (v(t), w(t)) be a trajectory of system (1) which goes through a point P of the half-plane v > 0, by (5) and the denition of the function U (l), the trajectories of the system (3) either intersect the integral curves of equation (6) in the direction of the origin or intersect the half-axis w < 0. Assuming the opposite, suppose that the trajectory lies completely inside the bounded region limited by axis w and some trajectory L1 of (6). We write the system (3) in the form: v w = (y + (1 ))w f (x + (1 ))v, = a(t)g (x + (1 ))v (7)
where 0 , , 1. x(t) = v(t) + (t) and y(t) = w(t) + (t) is a bounded solution, from this we have: f (x + (1 )) A < +, g (x + (1 )) > 0, for some positive constants A and . n if w > 0 N if w < 0 Then if v(t) v0 > 0 for all t t0 we have that w (t) av0 and w(t) , as t +. But this is a contradiction. So: Since (y) is bounded we obtain v Av + Dw where: D = w1 < w0 From this it follows that the origin is globally stable for the system (3). (8)
Examples and related remarks

0 < g (x) < + For U (l) we can take: U (l) = (a)/(nl r) (A)/(nl r) if l > a/n , if l a/n (9)
Example 1. Let:
when a(t) 1 and (y) = y, the condition (8) for the integral curves of equation (6) coincides with the convergence condition in (6). Example 2. When f (x) r, the convergence criterion for system (1) obtained in Example 1 can be improved if, in the choice of U (l), we take into account not only the greatest and least values of g (x) and f (x) but also relations among them. Apparently the rst criterion of this type was obtained in [5]. If we consider the case where an estimate: 0 < g (x) s + tf (x) < +; s, t R+ , holds. Then we have: (a)(ln r) U (l) = ((s + tr)A)(nl r) t
if l > r/n if r/n > l > s/tn if l s/tn
j. e. napoles a. i. ruiz
If we set (y) = y, a(t) 1 and = 0, we obtain: if l > r 0 (s + t)/(l r) if r > l > s/t, U (l) = t if l < s/t which coincides with the results given in [2]. Example 3. We take now: 0 < g (x) s + t thus we derive that: a/(nl r) (s + t r)A/(ln r) U (l) = 2 t /2(s s2 t2 nl) f (x) < +, t, s R+ ,
if l > r/n if r/n > l > r/n 2s r/tn if l r/n 2s r/tn
When a(t) 1, this result coincides with the one obtained in [8]. Example 4. We consider a forced oscillator with diferential equation: x + cx + g(x) = p(t), (10)
where c > 0 is a xed constant, g : R R is suciently smooth and has limit at innity, satisfying the inequality g() < g(x) < g(+), forall x R. The function p L (R) is bounded but not necessarily periodic. We remark that system (1) is the equation (10) under suitable conditions. Assume that 0 < g (x) < c2 /4, our Theorem 1 coincides with Theorem 1.2 in [1]. Remark. If we take g (x) and f (x) bounded functions, our results are consistent with those obtained in [7].
References
[1] Alonso, J.M.; Ortega R. (1995) Boundedness and global asymptotic stability of a forced oscillator, Nonlinear Anal. 25: 297309. [2] Bibikov, Y.N. (1976) Convergence in Linard equation with a forcing term, Vestnik LGU 7: 7375 e (russian). [3] Npoles, J.E. On the boundedness and global stability of solutions of a system of dierential equations a , to appear in Rev. Ciencias Matemticas, Universidad de La Habana (spanish). a [4] Npoles, J.E. On the global stability of non-autonomous systems, submited for publication. a [5] Nazarov, E.A. (1977) The coming together of the solutions of Linards equation , Dierencial nye e Uravnenija 13: 17921795 (russian). [6] Opial, Z. (1960-61) Sur un thor`me de C.E. Langenhop et G. Seifert , Ann. Polon. Math. 9: 145155. e e [7] Repilado, J.A.; J.E. Npoles. On the convergence of solutions of a bidimensional system to unique a periodic solution , submited for publication. [8] Ruiz, A. (1993) On the convergence of solutions of system x = h(y) f (x), y = g(x) + p(t) to a bounded solution , Rev. Ciencias Matemticas, Universidad de La Habana (spanish). a

Convergence in Nonlinear Systems With A Forcing Term: 1 Preliminars

Uploaded by

Copyright:

Available Formats

Convergence in Nonlinear Systems With A Forcing Term: 1 Preliminars

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Convergence in Nonlinear Systems With A Forcing Term: 1 Preliminars

Uploaded by

Copyright:

Available Formats

Revista de Matematica: Teor y Aplicaciones 4(1): 14 (1997) a

convergence in nonlinear systems with a forcing term

We consider the system:

a(t)g (s) kf (s) ds = v a(t)g (u) kf (u) + kl (z) ,

a(t)g (u) := R(l, t, u, z). l (z) f (u)

convergence in nonlinear systems with a forcing term

Examples and related remarks

if l > r/n if r/n > l > s/tn if l s/tn

if l > r/n if r/n > l > r/n 2s r/tn if l r/n 2s r/tn

You might also like