Surveys in Mathematics and its Applications ISSN 1842-6298 (electronic), 1843-7265 (print) Volume 7 (2012), 49 – 68 GLOBAL EXISTENCE FOR VOLTERRA–FREDHOLM TYPE NEUTRAL IMPULSIVE FUNCTIONAL INTEGRODIFFERENTIAL EQUATIONS V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan Abstract. In this paper, we study the global existence of solutions for the initial value problems for Volterra-Fredholm type neutral impulsive functional integrodifferential equations. Using the Leray-Schauder’s Alternative theorem, we derive conditions under which a solution exists globally. An application is provided to illustrate the theory. 1 Introduction In recent years, a great deal of work has been done in the study of the existence of solutions for impulsive differential equations, by which a number of chemotherapy, population dynamics, optimal control, industrial robotics and physics phenomena are described. For the general aspects of impulsive differential equations, we refer the reader to the classical monographs Bainov et al. [1, 2], Lakshmikantham et al. [20], Samoilenko and Perestyuk [26] and Hernández et al. [12, 11] for the case of ordinary and partial differential functional differential equations with impulses. For some general and recent works on the theory of impulsive differential and integrodifferential equations, we refer the reader to [3, 4, 5, 6, 8, 7, 9, 19, 31]. Recently, in [21], Ntouyas studied the global existence for first order neutral impulsive functional integrodifferential equations of the form d [x(t) − g(t, xt )] = f t, xt , dt x0 = φ Z t k(t, s)h(s, xs )ds , t ∈ I = [0, T ], 0 and also studied the global existence for second order neutral impulsive functional 2010 Mathematics Subject Classification: 34A60, 34K05, 34K10, 45N05, 45J05. Keywords: A priori bounds; Neutral impulsive functional integrodifferential equations; Global existence. ****************************************************************************** http://www.utgjiu.ro/math/sma 50 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan integrodifferential equations of the form Z t d 0 [x (t) − g(t, xt )] = F t, xt , k(t, s)H(s, xs , x0 (s))ds, x0 (t) , t ∈ I = [0, T ], dt 0 x0 = φ, x0 (0) = η by using Leray-Schauder’s Alternative theorem. Recent results on global existence and global solutions for ordinary and partial neutral impulsive integrodifferential equations can be foun in [13, 14, 15, 16, 17, 18, 22, 24, 23, 25, 27, 28, 29, 30]. In this paper, we study the global existence of solutions for the initial value problems for the first and second order Volterra-Fredholm type neutral impulsive functional integrodifferential equations. In Section 3, we study the following initial value problem d [x(t) − g(t, xt )] = f t, xt , dt Z t Z a(t, s)w(s, xs )ds, T b(t, s)h(s, xs )ds , 0 0 t ∈ I = [0, T ] \ {t1 , ..., tm }, ∆x|t=tk = Ik (x(t− k )), k = 1, ..., m, x0 = φ (1.1) (1.2) (1.3) where f : I × D × Rn × Rn → Rn , g : I × D → Rn , D = {ψ : [−r, 0] → Rn ; ψ is continuous everywhere except for a finite number of points e t at which ψ(e t− ) and ψ(e t+ ) are exist with ψ(e t− ) = ψ(e t)}, w : I × D → Rn , h : I × D → Rn , a : I × I → R, b : I × I → R are continuous functions, φ ∈ D, 0 < r < ∞, 0 = t0 < t1 < · · · < tm < + tm+1 = T , Ik ∈ C(Rn , Rn ), (k = 1, ..., m). x(t− k ) and x(tk ) represent the left and − right limits of x(t) at t = tk , respectively, ∆x|t=tk = x(t+ k ) − x(tk ). In Section 4, we study the global existence of solutions for the initial value problems for the second order Volterra-Fredholm type neutral impulsive functional integrodifferential equations of the form Z t d 0 0 [x (t) − g(t, xt )] = F t, xt , x (t), a(t, s)W (s, xs , x0 (s))ds, dt 0 Z T 0 b(t, s)H(s, xs , x (s))ds , t ∈ I = [0, T ] \ {t1 , ..., tm }, (1.4) 0 ∆x|t=tk = Ik (x(t− k )), 0 ∆x |t=tk = Jk (x(t− k )), 0 k = 1, ..., m, (1.5) k = 1, ..., m, and (1.6) x0 = φ, x (0) = η, (1.7) where F : I × D × Rn × Rn × Rn → Rn , g : I × D → Rn , D = {ψ : [−r, 0] → Rn ; ψ is continuous everywhere except for a finite number of points e t at which ψ(e t− ) and ψ(e t+ ) are exist with ψ(e t− ) = ψ(e t)}, W : I × D × Rn → Rn , H : I × D × Rn → Rn , ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma Global Existence For Volterra–Fredholm Type Functional Differential Equations 51 a : I × I → R, b : I × I → R are continuous functions, φ ∈ D and η ∈ Rn , 0 < r < ∞, 0 = t0 < t1 < · · · < tm < tm+1 = T , Ik , Jk ∈ C(Rn , Rn ), (k = 1, ..., m). + x(t− k ) and x(tk ) represent the left and right limits of x(t) at t = tk , respectively, − 0 + 0 − 0 ∆x|t=tk = x(t+ k ) − x(tk ), and ∆x |t=tk = x (tk ) − x (tk ). For any continuous function x defined on [−r, T ] \ {t1 , ..., tm } and any t ∈ [0, T ] we denote by xt the element of D defined by xt (θ) = x(t + θ), θ ∈ [−r, 0], where xt (·) represents the history of the state from time t − r, up to the present time t. 2 Preliminaries In this section, we introduce some basic definitions, notations and preliminary facts which are used throughout this paper. Let C([−r, 0], Rn ) be the Banach space of continuous functions from [−r, 0] into Rn endowed with the norm kφk = sup |φ(0)| : −r ≤ θ ≤ 0 and C([0, T ], Rn ) denote the Banach space of all continuous functions from [0, T ] into Rn normed by kxkr = sup{|x(t)| : −r ≤ t ≤ T }, kxk0 = sup{|x(t) : t ∈ I}, kxk1 = sup{|x0 (t) : t ∈ I}, kxk∗ = max{kxkr , kxk1 }, kxkT = max{kxk0 , kxk1 }. In order to define the solutions of (1.1)-(1.3) and (1.4)-(1.7), we introduce the following space: n P C = x : [0, T ] → Rn :xk ∈ C([tk , tk+1 ], Rn ), k = 0, ..., m and there exist o + − + x(t− ) and x(t ) with x(t ) = x(t ), k = 1, ..., m k k k k which is a Banach space with the norm kxkP C = max{kxk k(tk ,tk+1 ] , k = 0, ..., m}, where xk is the restriction of x to (tk , tk+1 ], k = 0, ..., m. Set Ω = D ∪ P C. Then Ω is a Banach space normed by kxkΩ = max{kxkD , kxkP C }, for each x ∈ Ω. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 52 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan and n P C 1 = x : [0, T ] → Rn :xk ∈ C 1 ([tk , tk+1 ], Rn ), k = 0, ..., m and there exist 0 + 0 − 0 + x0 (t− k ) and x (tk ) with x (tk ) = x (tk ), k = 1, ..., m o which is a Banach space with the norm kxkP C 1 = max{kxk k(tk ,tk+1 ] , k = 0, ..., m}, where xk is the restriction of x to (tk , tk+1 ], k = 0, ..., m. Set Ω1 = D ∪ P C 1 . Then Ω1 is a Banach space normed by kxkΩ1 = max{kxkD , kxkP C 1 }, for each x ∈ Ω1 . The considerations of this paper are based on the following fixed point result [10]. Lemma 1 (Leray-Schauder’s Alternative Theorem). Let S be a closed convex subset of a normed linear space E and assume that 0 ∈ S. If F : S → S be a completely continuous operator, i.e. it is continuous and the image of any bounded set is included in a compact set and let Φ(F ) = {x ∈ S : x = λF x, for some 0 < λ < 1}. Then either Φ(F ) is unbounded or F has a fixed point. 3 Global Existence for First order IVP In this section, we present the global existence results for the IVP (Initial value problem) (1.1)-(1.3). Definition 2. A function x ∈ Ω is called solution of the initial value problem (1.1)-(1.3) if x satisfies the following integral equation Z t Z s x(t) = φ(0) − g(0, φ) + g(t, xt ) + f s, xs , a(s, σ)w(σ, xσ )dσ, 0 0 Z T X b(s, σ)h(σ, xσ )dσ ds + Ik (x(t− k )), t ∈ I. 0 0<tk <t Theorem 3. Let f : I × D × Rn × Rn → Rn , g : I × D → Rn , w : I × D → Rn , h : I × D → Rn , a : I × I → R, b : I × I → R are continuous functions. Assume that ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma Global Existence For Volterra–Fredholm Type Functional Differential Equations 53 Hg1 There exists constants c1 and c2 such that |g(t, φ)| ≤ c1 kφk + c2 , t ∈ I, φ ∈ D. Hg2 The function g is completely continuous and such that the operator G : D → P C([0, T ], Rn ) defined by (Gφ)(t) = g(t, φ) is compact. Hw There exists a continuous function m1 : I → [0, ∞) such that |w(t, φ)| ≤ m1 (t)Ω1 (kφk), t ∈ I, φ ∈ D, where Ω1 : [0, ∞) → (0, ∞) is a continuous nondecreasing function. Hh There exists a continuous function m2 : I → [0, ∞) such that |h(t, φ)| ≤ m2 (t)Ω2 (kφk), t ∈ I, φ ∈ D, where Ω2 : [0, ∞) → (0, ∞) is a continuous nondecreasing function. Hf There exists an integrable function p : I → [0, ∞) such that |f (t, u, v, w)| ≤ p(t)Ω3 (kuk + |v| + |w|), t ∈ I, u ∈ D, v, w ∈ Rn , where Ω3 : [0, ∞) → (0, ∞) is a continuous nondecreasing function. Ha There exists a constant L1 such that |a(t, s)| ≤ L1 for t ≥ s ≥ 0. Hb There exists a constant L2 such that |b(t, s)| ≤ L2 for t ≥ s ≥ 0. HI There exist constants ck such that |Ik (x)| ≤ ck |x|, k = 1, ..., m for each x ∈ Rn . Then if Z 0 T Z m(s)ds b < c +∞ ds Ω3 (s) + Ω1 (s) + Ω2 (s) where m(t) b = max c= 1 (1 − c1 ) o 1 p(t), L1 m1 (t), L2 m2 (t) and (1 − c1 ) n n o X (1 + c1 )kφk + 2c2 + ck |x(t− )| , k n k=1 the initial value problem (1.1)-(1.3) has at least one solution on [−r, T ]. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 54 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan Proof. To prove the existence of a solution of the initial value problem (1.1)-(1.3) we apply Lemma 1. First we obtain the a priori bounds for the solutions of the initial value problem (1.1)λ − (1.3), λ ∈ (0, 1), where (1.1)λ stands for the equation Z T Z t d b(t, s)h(s, xs )ds , t ∈ I. a(t, s)w(s, xs )ds, [x(t) − g(t, xt )] = λf t, xt , dt 0 0 Let x be a solution of the initial value problem (1.1)λ − (1.3). From Z t Z s x(t) = λφ(0) − λg(0, φ) + λg(t, xt ) + λ f s, xs , a(s, σ)w(σ, xσ )dσ, 0 0 Z T X b(s, σ)h(σ, xσ )dσ ds + λ Ik (x(t− k )), t ∈ I. 0 0<tk <t we have, for every t ∈ I, t Z Z 0 0 T Z + L2 0 s m1 (σ)Ω1 (kxσ k)dσ p(s)Ω3 kxs k + L1 |x(t)| ≤ kφk + c1 kφk + 2c2 + c1 kxt k + n X ck |x(t− m2 (σ)Ω2 (kxσ k)dσ ds + k )|. k=1 We consider the function µ given by µ(t) = sup{|x(s)| : −r ≤ s ≤ t}, t ∈ I. Let t∗ ∈ [−r, t] be such that µ(t) = |x(t∗ )|. If t∗ ∈ [0, t], by the previous inequality we have, for every t ∈ I, Z t Z s µ(t) ≤ (1 + c1 )kφk + 2c2 + c1 µ(t) + p(s)Ω3 µ(s) + L1 m1 (σ)Ω1 (µ(σ))dσ 0 Z + L2 T 0 m2 (σ)Ω2 (µ(σ))dσ ds + 0 n X ck |x(t− k )| k=1 or Z t Z s n 1 µ(t) ≤ (1 + c1 )kφk + 2c2 + p(s)Ω3 µ(s) + L1 m1 (σ)Ω1 (µ(σ))dσ (1 − c1 ) 0 0 Z T n o X + L2 m2 (σ)Ω2 (µ(σ))dσ ds + ck |x(t− )| k 0 If k=1 t∗ ∈ [−r, 0] then µ(t) = kφk and the previous inequality obvious holds. Denoting by u(t) the right-hand side of the above inequality we have, u(0) = n n o X 1 (1 + c1 )kφk + 2c2 + ck |x(t− )| ≡ c, µ(t) ≤ u(t), t ∈ I, k (1 − c1 ) k=1 ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma Global Existence For Volterra–Fredholm Type Functional Differential Equations 55 and for every t ∈ I Z t 1 m1 (σ)Ω1 (µ(σ))dσ u (t) = p(t)Ω3 µ(t) + L1 (1 − c1 ) 0 Z T m2 (σ)Ω2 (µ(σ))dσ + L2 0 Z t 1 ≤ m1 (σ)Ω1 (u(σ))dσ p(t)Ω3 u(t) + L1 (1 − c1 ) 0 Z T m2 (σ)Ω2 (u(σ))dσ . + L2 0 0 Set t Z Z T m2 (s)Ω2 (u(s))ds, m1 (s)Ω1 (u(s))ds + L2 v(t) = u(t) + L1 t ∈ I. 0 0 Then v(0) = u(0) ≡ c, u(t) ≤ v(t), t ∈ I, u0 (t) ≤ 1 p(t)Ω3 (v(t)), t ∈ I (1 − c1 ) and v 0 (t) = u0 (t) + L1 m1 (t)Ω1 (u(t)) + L2 m2 (t)Ω2 (u(t)) 1 ≤ p(t)Ω3 (v(t)) + L1 m1 (t)Ω1 (v(t)) + L2 m2 (t)Ω2 (v(t)) (1 − c1 ) ≤ m(t) b Ω3 (v(t)) + Ω1 (v(t)) + Ω2 (v(t)) , t ∈ I. or v 0 (t) ≤ m(t), e t ∈ I. Ω3 (v(t)) + Ω1 (v(t)) + Ω2 (v(t)) Integrating this inequality from 0 to t and using the change of variables formula we get Z v(t) v(0) ds ≤ Ω3 (s) + Ω1 (s) + Ω2 (s) Z 0 T Z m(s)ds b < c +∞ ds , t ∈ I. Ω3 (s) + Ω1 (s) + Ω2 (s) This inequality implies that there is a constant K such that u(t) ≤ K, t ∈ I and hence µ(t) ≤ K, t ∈ I. Since for every t ∈ I, kxt k ≤ µ(t) we have kxkr = sup{|x(t)| : −r ≤ t ≤ T } ≤ K, ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 56 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan where K depends only on the functions m1 , m2 , p, Ω1 , Ω2 and Ω3 . We will rewrite the initial value problem (1.1)-(1.3) as follows. For φ ∈ D define φe ∈ B, B = Ω by ( φ(t), −r ≤ t ≤ 0, e φ= φ(0), 0 ≤ t ≤ T. e If x(t) = y(t) + φ(t), t ∈ [−r, T ] it is easy to see that y satisfies y0 = 0 Z t Z s e e y(t) = −g(0, φ) + g(t, yt + φt ) + f s, ys + φs , a(s, τ )w(τ, yτ + φeτ )dτ, 0 0 Z T X b(s, τ )h(τ, yτ + φeτ )dτ ds + Ik (x(t− k )), t ∈ I. 0 0<tk <t Define B0 = {y ∈ B : y0 = 0} and S : B0 → B0 , by Now define an operator Q : E0 → E by 0, t ∈ [−r, 0] Rs Rt e f s, ys + φes , 0 a(s, τ )w(τ, yτ + φeτ )dτ, −g(0, φ) + g(t, y + φ ) + t t (Sy)(t) = 0 R T b(s, τ )h(τ, y + φe )dτ ds + P t ∈ I. Ik (x(t− )), τ τ 0<tk <t 0 k S is clearly continuous. We shall prove that S is completely continuous. By Hg2 it suffices to prove that the operator S1 : B0 → B0 defined by 0, t ∈ [−r, 0] Rt R e s yτ + φeτ )dτ, (S1 y)(t) = 0 f τ, ys + φs , 0 a(s, τ )w(τ, R T b(s, τ )h(τ, y + φe )dτ ds, t ∈ I. 0 τ τ is completely continuous. Let {uv } be a bounded sequence in B0 , i.e. kuv kr ≤ b, for all v, where b is a positive constant. We obviously have kuvt k ≤ b, t ∈ I, for all v. Hence, if M ∗ = sup{m e : t ∈ I}, p0 = sup{p(t) : t ∈ I}, we obtain h i M ∗T 2 M ∗T 2 kS1 uv kr ≤ p0 Ω3 (b + kφk)T + Ω1 (b + kφk) + Ω2 (b + kφk) . 2 2 This means that {S1 uv } is uniformly bounded. Moreover, the sequence {S1 uv } is equicontinuous, since for t1 , t2 ∈ [−r, T ] we have: ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma Global Existence For Volterra–Fredholm Type Functional Differential Equations 57 (i) if 0 ≤ t1 ≤ t2 then |S1 uv (t1 ) − S1 uv (t2 )| ≤ p0 Ω3 (b + kφk) + M ∗ T (Ω1 (b + kφk) + Ω2 (b + kφk) |t1 − t2 |. (ii) if t1 ≤ 0 ≤ t2 then |S1 uv (t1 ) − S1 uv (t2 )| ≤ p0 Ω3 (b + kφk) + M ∗ T (Ω1 (b + kφk) + Ω2 (b + kφk) |0 − t2 |. (iii) if t1 ≤ t2 ≤ 0 then |S1 uv (t1 ) − S1 uv (t2 )| = 0. Thus, by Arzela-Ascoli theorem the operator S1 is completely continuous. Consequently the operator S is completely continuous. Finally, the set Φ(S) = {y ∈ B0 : y = λSy, λ ∈ (0, 1)} is bounded, since for every solution y ∈ B0 the function x = y + φ is a solution of the initial value problem (1.1)-(1.3), for which we have proved that kxkr ≤ K and hence kykr ≤ K + kφk. Consequently, by Lemma 1, the operator S has a fixed point y ∗ in B0 . Then ∗ x = y ∗ + φe is a solution of the initial value problem (1.1)-(1.3). Hence the proof of the theorem is complete. 4 Global Existence for Second order IVP In this section we study the global existence for the IVP (Initial value problem) (1.4)-(1.7). Definition 4. A function x ∈ Ω ∩ Ω1 is called solution of the initial value problem (1.4)-(1.7) if x satisfies the following integral equation Z t Z tZ s x(t) = φ(0) + [η − g(0, φ)]t + g(s, xs )ds + F τ, xτ , x0 (τ ), 0 0 0 Z τ Z T a(τ, σ)W (σ, xσ , x0 (σ))dσ, b(τ, σ)H(σ, xσ , x0 (σ))dσ dτ ds 0 0 X − + [Ik (x(tk )) + (T − tk )Jk (x(t− k ))], t ∈ I. 0<tk <t Theorem 5. Let F : I ×D ×Rn ×Rn ×Rn → Rn , g : I ×D → Rn , W : I ×D ×Rn → Rn , H : I × D × Rn → Rn are continuous functions and a : I × I → R, b : I × I → R are measurable for t ≥ s ≥ 0 function. Assume that HI is satisfied and g satisfies Hg1 , Hg2 , a, b satisfies Ha and Hb . Moreover we assume that: ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 58 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan HW There exists a continuous function m3 : I → [0, ∞) such that |W (t, φ, ψ)| ≤ m3 (t)Ω4 (kφk + |ψ|), t ∈ I, φ ∈ D, ψ ∈ Rn , where Ω4 : [0, ∞) → (0, ∞) is a continuous nondecreasing function. HH There exists a continuous function m4 : I → [0, ∞) such that |H(t, φ, ψ)| ≤ m4 (t)Ω5 (kφk + |ψ|), t ∈ I, φ ∈ D, ψ ∈ Rn , where Ω5 : [0, ∞) → (0, ∞) is a continuous nondecreasing function. HF There exists an integrable function P : I → [0, ∞) such that |F (t, u, v, w, y)| ≤ P (t)Ω6 (kuk + |v| + |w| + |y|), t ∈ I, u ∈ D, v, w, y ∈ Rn , where Ω6 : [0, ∞) → (0, ∞) is a continuous nondecreasing function. HJ There exist constants dk such that |Jk (x)| ≤ dk |x|, k = 1, ..., m for each x ∈ Rn . Then if Z T Z ∞ c(s)ds < M 0 c0 ds s + 2Ω6 (s) + Ω4 (s) + Ω5 (s) where Z t n o c(t) = max (1 + c1 )c1 , (1 + c1 ) M P (τ )dτ, P (t), L1 m3 (t), L2 m4 (t) , and 0 c0 = kφk + |η| + c1 kφk + 2c2 ) (1 + T ) + c1 u(0) n X − ck |x(t− + k )| + (T − tk )dk |x(tk )| k=1 the initial value problem (1.4)-(1.7) has at least one solution on [−r, T ]. Proof. To prove the existence of a solution of the initial value problem (1.4)-(1.7) we apply Lemma 1. First we obtain the a priori bounds for the solutions of the initial value problem (1.4)λ − (1.7), λ ∈ (0, 1), where (1.4)λ stands for the equation Z t d 0 [x (t) − g(t, xt )] =λF t, xt , x0 (t), a(t, s)g(s, xs , x0 (s))ds, dt 0 Z T b(t, s)h(s, xs , x0 (s))ds , t ∈ I. 0 ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma Global Existence For Volterra–Fredholm Type Functional Differential Equations 59 Let x be a solution of the initial value problem (1.4)λ − (1.7). From Z Z tZ t s F τ, xτ , x0 (τ ), g(s, xs )ds + λ x(t) = λφ(0) + λ[η − g(0, φ)]t + λ 0 0 0 Z T Z τ b(τ, σ)h(σ, xσ , x0 (σ))dσ dτ ds a(τ, σ)w(σ, xσ , x0 (σ))dσ, 0 0 X − +λ [Ik (x(tk )) + (T − tk )Jk (x(t− k ))], t ∈ I 0<tk <t we have, for every t ∈ I, Z tZ s Z t P (τ )Ω6 kxτ k + |x0 (τ )| kxs kds + |x(t)| ≤ kφk + [|η| + c1 kφk + 2c2 ]T + c1 0 0 0 Z T Z τ m4 (σ)Ω5 (kxσ k)dσ dτ ds m3 (σ)Ω4 (kxσ k)dσ + L2 + L1 0 0 + n X − [ck |x(t− k )| + (T − tk )dk |x(tk )|]. k=1 We consider the function µ given in the proof of Theorem 3. Then by the previous inequality we have, µ(t) = sup{|x(s)| : −r ≤ s ≤ t}, t ∈ I. Let t∗ ∈ [−r, t] be such that µ(t) = |x(t∗ )|. If t∗ ∈ [0, t], by the previous inequality we have, for every t ∈ I, |µ(t)| Z t Z tZ s ≤ kφk + [|η| + c1 kφk + 2c2 ]T + c1 µ(s)ds + P (τ )Ω6 µ(τ ) + |x0 (τ )| 0 0 0 Z T Z τ m4 (σ)Ω5 (µ(σ) + |x0 (σ)|)dσ dτ ds + L1 m3 (σ)Ω4 (µ(σ) + |x0 (σ)|)dσ + L2 0 + n X 0 − [ck |x(t− k )| + (T − tk )dk |x(tk )|]. k=1 If t∗ ∈ [−r, 0] then µ(t) = kφk and the previous inequality obvious holds. Denoting by u(t) the right hand side of the above inequality we have, n X − u(0) = kφk + [|η| + c1 kφk + 2c2 ]T + [ck |x(t− k )|+(T − tk )dk |x(tk )|], k=1 µ(t) ≤ u(t), t ∈ I. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 60 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan and Z τ Z t m3 (σ)Ω4 (µ(σ) + |x0 (σ)|)dσ P (τ )Ω6 µ(τ ) + |x0 (τ )| + L1 u0 (t) = c1 µ(t) + 0 0 Z T m4 (σ)Ω5 (µ(σ) + |x0 (σ)|)dσ dτ + L2 0 Z τ Z t 0 m3 (σ)Ω4 (u(σ) + |x0 (σ)|)dσ P (τ )Ω6 u(τ ) + |x (τ )| + L1 ≤ c1 u(t) + 0 0 Z T m4 (σ)Ω5 (u(σ) + |x0 (σ)|)dσ dτ, t ∈ I. + L2 0 Therefore if v(t) = sup{|x0 (s)| : s ∈ I}, t ∈ I we obtain Z 0 t Z τ P (τ )Ω6 u(τ ) + v(τ ) + L1 m3 (σ)Ω4 (u(σ) + v(σ))dσ u (t) ≤ c1 u(t) + 0 0 Z T + L2 m4 (σ)Ω5 (u(σ) + v(σ))dσ dτ, t ∈ I. (4.1) 0 On the other hand, by t x (t) =λ[η − g(0, φ)] + λg(t, xt ) + λ F τ, xτ , x0 (τ ), 0 Z τ Z T 0 a(τ, s)W (s, xs , x (s))ds, b(τ, s)H(s, xs , x0 (s))ds dτ Z 0 0 0 for any t ∈ I and every s ∈ [0, t], we obtain Z t Z τ 0 0 |x (t)| ≤|η| + c1 kφk + 2c2 + c1 kxt k + P (τ )Ω6 u(τ ) + |x (τ )| + L1 m3 (σ) 0 0 Z T (×)Ω4 (u(σ) + |x0 (τ )|)dσ + L2 m4 (σ)Ω5 (u(σ) + |x0 (τ )|)dσ dτ, 0 or Z t Z τ v(t) ≤|η| + c1 kφk + 2c2 + c1 u(t) + P (τ )Ω6 u(τ ) + v(τ ) + L1 m3 (σ) 0 0 Z T (×)Ω4 (u(σ) + v(τ ))dσ + L2 m4 (σ)Ω5 (u(σ) + v(τ ))dσ dτ, t ∈ I. 0 Denoting by z(t) the right hand side of the above inequality we have: z(0) = |η| + c1 kφk + 2c2 + c1 u(0), v(t) ≤ z(t), t ∈ I ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma Global Existence For Volterra–Fredholm Type Functional Differential Equations 61 and Z t m3 (σ)Ω4 (u(σ) + v(τ ))dσ z 0 (t) = c1 u0 (t) + P (t)Ω6 u(t) + v(t) + L1 0 Z T m4 (σ)Ω5 (u(σ) + v(τ ))dσ , + L2 0 Z t 0 m3 (σ)Ω4 (u(σ) + z(τ ))dσ ≤ c1 u (t) + P (t)Ω6 u(t) + z(t) + L1 0 Z T m4 (σ)Ω5 (u(σ) + z(τ ))dσ , t ∈ I. + L2 0 From (4.1), since v(t) ≤ z(t) we have Z τ Z t m3 (σ)Ω4 (u(σ) + z(σ))dσ u0 (t) ≤ c1 u(t) + P (τ )Ω6 u(τ ) + z(τ ) + L1 0 0 Z T + L2 m4 (σ)Ω5 (u(σ) + z(σ))dσ dτ, t ∈ I. 0 Let Z t w(t) = u(t) + z(t)+L1 m1 (σ)Ω4 (u(σ) + z(σ))dσ Z T + L2 m2 (σ)Ω5 (u(σ) + z(σ))dσ, 0 t ∈ I. 0 Then u(t) + z(t) ≤ w(t), t ∈ I w(0) = u(0) + z(0) = c0 , and w0 (t) = u0 (t) + z 0 (t) + L1 m3 (t)Ω4 (u(t) + z(t)) + L2 m4 (t)Ω5 (u(t) + z(t)) Z t ≤ c1 u(t) + P (τ )Ω6 (w(τ ))dτ + c1 u0 (t) + P (t)Ω6 (w(t)) 0 + L1 m3 (t)Ω4 (u(t) + z(t)) + L2 m4 (t)Ω5 (u(t) + z(t)) Z t ≤ (1 + c1 )c1 w(t) + (1 + c1 ) P (τ )Ω6 (w(τ ))dτ + P (t)Ω6 (w(t)) 0 + L1 m3 (t)Ω4 (u(t) + z(t)) + L2 m4 (t)Ω5 (u(t) + z(t)) ≤ m(t) b w(t) + 2Ω6 (w(t)) + Ω4 (w(t)) + Ω5 (w(t)) , t ∈ I. This implies Z w(t) w(0) ds ≤ s + 2Ω6 (s) + Ω4 (s) + Ω5 (s) Z T Z ∞ c(s)ds < M 0 c0 ds s + 2Ω6 (s) + Ω4 (s) + Ω5 (s) ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 62 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan This inequality implies that there is a constant K such that w(t) ≤ K, t ∈ I. Then |x(t)| ≤ µ(t) ≤ u(t), t ∈ I |x0 (t)| ≤ v(t) ≤ z(t), t ∈ I and hence kx∗ k ≤ kxkr + kxk1 ≤ K. In the second step we rewrite the initial value problem (1.4)-(1.7) as an integral operator and will prove that this operator is completely continuous. Define the operator Q : B → B, B = Ω by φ(t), t ∈ [−r, 0] R R R φ(0) + [η − g(0, φ)]t + t g(s, ys )ds + t s F τ, xτ , x0 (τ ), 0 0 0 (Qx)(t) = Rτ RT 0 0 (σ))dτ dτ ds a(s, σ)W (σ, x , x (σ))dσ, b(s, σ)H(τ, x , x σ σ 0 0 +P − − t ∈ I. 0<tk <t [Ik (x(tk )) + (T − tk )Jk (x(tk ))], Let a bounded sequence {zυ } in B, i.e kzυ k ≤ b, for all υ. Then the sequence {zυ (t)}, t ∈ I is bounded in B. Using the fact that Z t Z 0 0 a(t, s)W (s, zυs , zυ (s))ds, F t, zυt , zυ (s), 0 0 T b(t, s)H(s, zυs , zυ0 (s))ds ≤ M1 where f(t), t ∈ I}} M1 = M ∗ Ω6 2b + M ∗ Ω4 (2b)T + M ∗ Ω5 (2b)T , M ∗ = max{kφk, max{M we can easily prove that there exists a constant n T2 M = max kφk + [|η| + c1 kφk + 2c2 ]T + c1 bT + M1 2 n o X + (ck + (T − tk )dk )|z(t− )|, |η| + (1 + T )c kφk + (2 + T )c + M T 1 2 1 k k=1 such that kQzυ kr ≤ M and k(Qzυ )0 k1 ≤ M which means that {Qzυ } and {(Qzυ )0 } are uniformly bounded. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 63 Global Existence For Volterra–Fredholm Type Functional Differential Equations Moreover rewriting the operator Qx for t ∈ I as (Qx)(t) = (Q1 x)(t) + (Q2 x)(t) where Z t g(s, ys )ds (Q1 x)(t) =φ(0) + [η − g(0, φ)]t + 0 X − + [Ik (x(t− k )) + (T − tk )Jk (x(tk ))], Z tZ (Q2 x)(t) = 0 s F τ, xτ , x0 (τ ), Z 0<tk <t τ 0 0 a(s, σ)W (σ, xσ )dσ, Z T b(s, σ)H(τ, xσ )dτ dτ ds, t ∈ I, 0 it is easy to see that the sequences {Q1 zυ }, {(Q1 zυ )0 } and {Q2 zυ }, {(Q2 zυ )0 } are equicontinuous. Thus, by the Arzela-Ascoli theorem the operator Q is completely continuous. Finally, the set Φ(Q) = {x ∈ B : x = λQx, λ ∈ (0, 1)} is bounded, as we proved in the first part. Hence by Lemma 1, the operator Q has a fixed point in B. Then it is clear that the initial value problem (1.4)-(1.7) has at least one solution. Hence the proof of the theorem is complete. 5 Application In this section, we apply some of the results established in this paper. Example 6. First, we consider the partial first order differential equation of the form t Z wt (u, t) − g1 (t, w(x, t − r)) =P t, w(u, t), 0 k1 t, w(x, t − r) ds, Z b h1 t, w(x, t − r) ds , (5.1) 0 w(0, t) =w(π, t) = 0, w(u, t) =φ(x, t), − w(t+ k , y)−w(tk , y) 0 ≤ t ≤ b, (5.2) 0 ≤ u ≤ π, = Ik (w(t− k , y)), (5.3) k = 1, 2, ..., m, (5.4) where P : [0, b] × D × Rn × Rn → Rn is a function and g1 , k1 , h1 : [0, b] × D → Rn are continuous functions. We assume that the functions P, g1 , k1 and h1 in (5.1)-(5.4) satisfy the following conditions. (1) There exist constants c1 and c2 such that |g(t, φ)| ≤ c1 |φ| + c2 , for t ∈ [0, b], φ ∈ D. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 64 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan (2) There exists a nonnegative function p1 defined on [0, b] such that Z t k1 (t, x)ds ≤ p1 (t)|x| 0 for t ∈ [0, b] and x ∈ D. (3) There exists a nonnegative function q1 defined on [0, b] such that Z b h1 (t, x)ds ≤ q1 (t)|x| 0 for t ∈ [0, b] and x ∈ D. (4) There exists nonnegative real valued continuous function l1 defined on [0, b] and a positive continuous increasing function K1 defined on R+ such that |P (t, x, y, z)| ≤ l1 (t)K1 (|x| + |y| + |z|) for t, x, y, z ∈ [0, b] × D × Rn × Rn . (5) There exist constants ck such that |Ik (x)| ≤ ck |x|, k = 1, ..., m for each x ∈ Rn . Let us take X = L2 [0, π]. Suppose that Z 0 T Z m(s)ds b < +∞ c ds Ω3 (s) + Ω1 (s) + Ω2 (s) where m(t) b = max c= 1 (1 − c1 ) o 1 p(t), L1 m1 (t), L2 m2 (t) and (1 − c1 ) n n o X (1 + c1 )kφk + 2c2 + ck |x(t− )| k n k=1 is satisfied. Define the functions f : [0, b] × X × X × X → X, a, b : [0, b] × [0, b] → X, g, w, h : [0, b] × X → X, Ik ∈ (X, X) as follows f (t, x, y, z)(u) = P (t, x(u, t), y(u, t), z(u, t)), k(t, x)(u) = k1 (t, x(u, t)) and h(t, x)(u) = h1 (t, x(u, t)) for t ∈ [0, b], x, y, z ∈ X and 0 ≤ u ≤ π. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma Global Existence For Volterra–Fredholm Type Functional Differential Equations 65 With these choices of the functions, the equations (5.1)-(5.4) can be modelled abstractly as nonlinear mixed Volterra-Fredholm integrodifferential equation with nonlocal condition in Banach space X: d [x(t) − g(t, xt )] = f t, xt , dt Z t Z a(t, s)w(s, xs )ds, T b(t, s)h(s, xs )ds , 0 0 t ∈ I = [0, T ] \ {t1 , ..., tm }, ∆x|t=tk = Ik (x(t− k )), x0 = φ k = 1, ..., m, (5.5) (5.6) (5.7) Since all the hypotheses of the Theorem 3 are satisfied, the Theorem 3 can be applied to guarantee the solution of the nonlinear mixed Volterra-Fredholm type neutral impulsive integrodifferential equation. Acknowledgement. The first author would like to thank Dr. V. Palanisamy, Principal, Dr. Chitra Manohar, Secretary, Info Institute of Engineering, Coimbatore, for their constant encouragement and support for this research work. The third author dedicates this paper to “Silver Jubilee Year Celebrations of Karunya University, Coimbatore”. References [1] D.D. Bainov and P.S. Simeonov, Impulsive Differential Equations: Periodic Solutions and Applications, Longman Scientific and Technical Group, England, 1993. MR1266625(95b:34012). Zbl 0815.34001. [2] D.D. Bainov and V.C. Covachev, Impulsive Differential Equations with a Small Parameter, World Scientific, New Jersey, 1994. MR1430569(97i:34001). Zbl 0828.34001. [3] K. Balachandran, J.Y. Park and M. Chandrasekaran, Nonlocal Cauchy problem for delay integrodifferential equations of Sobolve type in Banach spaces, Appl. Math. Lett. 15(7) (2002), 845-854. MR1920985(2003d:34162). Zbl 1028.45006. [4] K. Balachandran, D.G. Park and S.M. Anthoni, Existence of solutions of abstract nonlinear second order neutral functional integrodifferential equations, Comput. Math. Appl. 46 (2003), 1313-1324. MR2019686(2005b:45006). Zbl 1054.45006. [5] M. Benchohra, J. Henderson and S.K. Ntouyas, Impulsive differential equations and inclusions, Hindawi Publishing Corporation, New York, 2006. xiv+366 pp. ISBN: 977-5945-50-X. MR2322133(2008f:34001). Zbl 1130.34003. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 66 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan [6] Y.K. Chang, V. Kavitha and M. Mallika Arjunan, Existence results for a second order impulsive functional differential equation with state-dependent delay, Diff. Equ. Appl. 1(3) (2009), 325-339. MR2554970(2010k:34223). Zbl 1192.34091. [7] Y.K. Chang, A. Anguraj and M. Mallika Arjunan, Existence results for impulsive neutral functional differential equations with infinite delay, Nonlinear Anal. Hybrid Syst. 2(1) (2008), 209-218. MR2382006(2008k:34027). Zbl 1170.35467. [8] Y.K. Chang, V. Kavitha and M. Mallika Arjunan, Existence results for impulsive neutral differential and integrodifferential equations with nonlocal conditions via fractional operators, Nonlinear Anal. Hybrid Syst. 4(1) (2010), 32-43. MR2570180(2010m:34151). Zbl 1200.34095. [9] M. B. Dhakne and S. D. Kendre, On abstract Nonlinear Mixed Volterra-Fredhom Integrodifferential equations, Comm. Appl. Nonlinear Anal. 13(4) (2006), 101112. MR2286408(2008d:45008). Zbl 1119.45005. [10] J. Dugundji and A. Granas, Fixed point theory, Vol. 1, Matematyczne, PNW Warsawa, 1982. MR0660439(83j:54038). Zbl 0483.47038. [11] E. Hernández, A second order impulsive Cauchy problem, Int. J. Math. Math. Sci. 31(8) (2002), 451-461. MR1930085(2003g:34021). Zbl 1013.34061. [12] E. Hernández, H.R. Henrı́quez and M.A. McKibben, Existence results for abstract impulsive second-order neutral functional differential equations, Nonlinear Anal. TMA, 70 (2009), 2736-2751. MR2499742(2010a:34190). Zbl 1173.34049. [13] E. Hernández and S.M. Tanaka Aki, Global solutions for abstract functional differential equations with nonlocal conditions, Electron. J. Qual. Theory Differ. Equ. 50 (2009), 1-8. MR2535075(2010i:34191). Zbl 1196.34102. [14] E. Hernández, S.M. Tanaka Aki and H.R. Henrı́quez, Global solutions for impulsive abstract partial differential equations, Comput. Math. Appl.56 (2008), 1206-1215. MR2437288(2009h:35432). Zbl 1155.35481. [15] E. Hernández, Global solutions for abstract impulsive neutral differential equations, Math. Comput. Model. 53 (2011), 196-204. MR2739257. Zbl 1211.34098. [16] E. Hernández and S.M. Tanaka Aki, Global solutions for abstract impulsive differential equations, Nonlinear Anal. TMA. 72 (2010), 1280-1290. MR2577529. Zbl 1183.34083. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma Global Existence For Volterra–Fredholm Type Functional Differential Equations 67 [17] E. Hernández and H.R. Henrı́quez, Global solutions for a functional second order abstract Cauchy problem with nonlocal conditions, Ann. Pol. Math. 83 (2004), 149-170. MR2111405(2005j:34087). Zbl 1113.47309. [18] E. Hernández, M. Rabello and H.R. Henrı́quez, Existence of solutions for impulsive partial neutral functional differential equations, J. Math. Anal. Appl. 331 (2007), 1135-1158. MR2313705(2008m:35356). Zbl 1123.34062. [19] V. Kavitha, M. Mallika Arjunan and C. Ravichandran, Existence results for impulsive systems with nonlocal conditions in Banach spaces, J. Nonlinear Sci. Appl. 4(2) (2011), 138-151. MR2783840. Zbl pre05902649. [20] V. Laksmikantham, D. Bainov and P.S. Simenov, Theory of impulsive differential equations, Series in Modern Applied Mathematics, 6. World Scientific Publishing Co., Inc., Teaneck, NJ, 1989. MR1082551(91m:34013. Zbl 0719.34002. [21] S.K. Ntouyas, Global existence for neutral functional integrodifferential equations, Nonlinear Anal. TMA, 30 (4) (1997), 2133-2142. MR1490335(99b:34132). Zbl 0890.45004. [22] S.K. Ntouyas and P. Tsamatos, Global existence for functional integrodifferential equations of delay and neutral type, Appl. Anal. 54 (1994), 251-262. MR1379636(96j:45005). Zbl 0838.34078. [23] S.K. Ntouyas and P. Tsamatos, Global existence for semilinear evolution integrodifferential equations with delay and nonlocal conditions, Appl. Anal. 64 , (1997), 99-105. MR1460074(98e:45008). Zbl 0874.35126. [24] S.K. Ntouyas and P. Tsamatos, Global existence for second order semilinear ordinary and delay integrodifferential equations with nonlocal conditions, Appl. Anal. 67 (1997), 245-257. MR1614061(98k:34103). Zbl 0906.35110. [25] B. Pachpatte, Applications of the Leray-Schauder Alternative to some Volterra integral and integrodifferential equations, Indian J. pure appl. Math. 26 (1995), 1161-1168. MR1364736(96m:45003). Zbl 0852.45012. [26] A.M. Samoilenko and N.A. Perestyuk; Impulsive Differential Equations, World Scientific, Singapore, 1995. MR1355787(97i:34002). Zbl 0837.34003. [27] S. Sivasankaran, M. Mallika Arjunan and V. Vijayakumar, Existence of global solutions for impulsive functional differential equations with nonlocal conditions, J. Nonlinear Sci. Appl. 4(2) (2011), 102-114. MR2783836. Zbl pre05902645. [28] S. Sivasankaran, M. Mallika Arjunan and V. Vijayakumar, Existence of global solutions for second order impulsive abstract partial differential equations, Nonlinear Anal. TMA. 74(17) (2011), 6747-6757. ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma 68 V. Vijayakumar, S. Sivasankaran and M. Mallika Arjunan [29] H.L. Tidke and M.B. Dhakne, Global existence of mild solutions of second order nonlinear Volterra integrodifferential quations in Banach spaces, Differ. Equ. Dyn. Syst. 17(4) (2009), 331-342. MR2610901(2011c:34198). Zbl 1206.35176. [30] H.L. Tidke, Existence of global solutions to nonlinear mixed volterra-fredholm integrodifferential equations with nonlocal conditions, Elect. J. Diff. Equ. 55 (2009), 1–7. MR2505113(2010b:34139). Zbl 1165.45010. [31] H.L. Tidke and M.B. Dhakne, Existence of solutions for nonlinear mixed type integrodifferential equation of second order, Surv. Math. Appl. 5 (2010), 61-72. MR2639727(2011g:34187). Zbl 1204.45015. V. Vijayakumar S. Sivasankaran Department of Mathematics, Department of Mathematics, Info Institute of Engineering, University College, Kovilpalayam, Coimbatore-641 107, Sungkyunkwan University, Tamilnadu, India. Suwon 440-746, South Korea. e-mail: vijaysarovel@gmail.com e-mail: sdsiva@gmail.com. M. Mallika Arjunan Department of Mathematics, Karunya University, Karunya Nagar, Coimbatore-641 114, Tamilnadu, India. e-mail: arjunphd07@yahoo.co.in ****************************************************************************** Surveys in Mathematics and its Applications 7 (2012), 49 – 68 http://www.utgjiu.ro/math/sma