Difference methods for impulsive differential-functional equations D r u m i D. B a i n o v a,., Z d z i s l a w K a m o n t b, E m i l M i n c h e v c a Southwestern University, Blagoevgrad, Bulgaria b University of Gdansk, Gdansk, Poland c Sofia University, Sofia, Bulgaria
Abstract
We consider a class of difference methods for initial value problems for first-order impulsive partial differentialfunctional equations. We give sufficient conditions for the convergence of a sequence of approximate solutions under the assumptions that the right-hand sides satisfy the nonlinear estimates of Perron type with respect to the functional argument. The proof of the stability of difference methods is based on a general theorem on the error estimate of approximate solutions for difference-functional equations of Volterra type with an unknown function of several variables.
1. I n t r o d u c t i o n
Many real processes and p h e n o m e n a studied in biology, mechanics and population dynamics are characterized by the fact that at certain moments of their development the system parameters undergo rapid changes. A natural tool for mathematical simulation of such processes and p h e n o m e n a is the theory of impulsive differential equations. The mathematical investigations of this theory mark their beginning with the work of V. Mil'man and A. Myshkis [14]. At first the theory of impulsive differential equations developed slowly, due to some difficulties of technical and theoretical character related to the presence of some characteristic peculiarities such as " b e a t i n g " , "dying", " m e r g i n g " , noncontinuability of the solutions, loss of the property of autonomy, etc. Recently, however, a considerable increase in the number of publications is observed in various branches of this theory [2-4,20]. The first works on impulsive partial differential equations were published after 1991 [1,6]. These are devoted to the qualitative theory of impulsive partial differential equations. It is
D.D. Bainov et al. /Applied Numerical Mathematics 16 (1995) 401-416
necessary to investigate the numerical methods for this new theory in order to use these works for mathematical simulation. In recent years a n u m b e r of papers concerned with difference methods for first-order partial differential equations [7,8,10,11,16,21] and for differential-functional equations [5,9,17,18,22] were published. A m e t h o d of difference inequalities and simple theorems on recurrent inequalities are used in the investigation of stability. Brandi et al. [5], Kamont and Przadka [8], Kowalski [10,11], Plis [16] and Przadka [17,18] have assumed that given functions have partial derivatives with respect to all arguments with the exception of (x, y). In [9] Kamont and Przadka initiated investigations of difference methods for partial equations with nonlinear estimates of Perron type with respect to the functional argument. In this paper we extend the results of [8,10,17,18] on the case of differential-functional equations with impulses. We consider a general class of difference schemes. The basic tool in the investigation of stability is a t h e o r e m on the error estimate for difference-functional equations of Volterra type. In this paper we use general ideas for finite difference approximations which were introduced in [12,13,15,19].
2. Preliminary notes For any two metric spaces X and Y we denote by C(X, Y) the class of all continuous functions from X into Y. We use the symbol E to denote the H a a r pyramid
i= 1 , . . . , n } , where a > 0, M/>1 0 and b i - M i a > 0 for i = 1 , . . . , n . Write b = ( b l , . . . , b , , ) , M = ( M 1 , . . . , M n) E = {(x, y) = (x, Yl , ' ' ' , Y n ) ~ ~l+n: X ~ [0, a], [Yi[ ~ b i - M i x ,
and E 0 = [ - r 0 , 0] × I - b , b] where r 0 ~ ~+, ~ + = [0, +oo). Suppose that 0 < a I < a 2 < a k < a are given numbers and a 0 --- 0, ak+ 1 = a. We define
--" <
E(i)={(x,y)~E:aiR such that (i) the functions z le,~, i = O, 1,..., k, and z lE° are continuous, (ii) for each i, 0 <~i <<.k, (ai, y) ~ E , there exists lim z(t, ~ ) = z ( a ~ , y), (t, ~) ---~(ai,Y) t>a i
(iii) for each i, 1 <~i <~k + 1, lim (t, ~)--*(ai,Y) t
(ai,