1 Introduction

Let (X,) be a Banach space, C[J,X]={x:J=[0,a]X,x(t) is continuous in J} with the norm x C = sup t J x(t). It is easy to verify that C[J,X] is a Banach space. The space of X-valued Bochner integrable functions on J with the norm x 1 = 0 a x(s)ds is denoted by L[J,X]. Consider the following nonlinear mixed type integro-differential functional evolution equations with nonlocal conditions in a Banach space X(IVP),

(1.1)
(1.2)

where

(K x σ 2 )(t)= 0 t k ( t , s , x ( σ 2 ( s ) ) ) ds,(H x σ 3 )(t)= 0 a h ( t , s , x ( σ 3 ( s ) ) ) ds,
(1.3)

A is the generator of a strongly continuous semigroup in the Banach space X, and F(t) is a bounded operator for tJ, x 0 X, fC[J× X 3 ,X], g:C[J,X]X, kC[×X,X], ={(t,s)J×J:st}, hC[J×J×X,X], σ i C[J,J] and σ i (t)t (i=1,2,3).

For the existence of mild solutions of integro-differential functional evolution equations in abstract spaces, there are many research results, see [116], and references therein. In order to obtain the existence and controllability of mild solutions in these study papers, usually, some restricted conditions on a priori estimation and compactness conditions of an evolution operator or compactness conditions on f(t, X r , X r , X r ) are used.

Recently, using a fixed point theorem, Haribhau Laxman Tidkey and Machindra Baburao Dhakne [1] have studied the existence of mild solutions of IVP (1.1)-(1.2) when σ i (t)=t (i=1,2,3), the compactness of the resolvent operator and the restricted condition

M 1 [ x 0 + G 1 + L r b + L K r b 2 + L K 1 b 2 + L H r b 2 + L H 1 b 2 + L 1 b ] r

with M 1 [Lb+LK b 2 +LH b 2 ]<1 is used. Malar [17] and Shi [18] studied the existence of mild solutions of semilinear mixed type integrodifferential evolution equations with the equicontinuous semigroup

{ x ( t ) = A x ( t ) + f ( t , x ( t ) , 0 t a ( t , s ) k ( s , x ( s ) ) d s , x ( t ) = 0 a b ( t , s ) h ( s , x ( s ) ) d s ) , t [ 0 , a ] , x ( 0 ) = x 0 + g ( x ) .
(1.4)

Solvability of the scalar equation

m(t)= K 1 + K 2 0 t h ( s , m ( s ) , n ( s ) , q ( s ) ) ds,tJ

and the restricted condition on measure of noncompactness estimation

0 t [ η 1 ( s ) + k 1 η 2 ( s ) + k 2 η 3 ( s ) ] dsK

are used in [17]. But estimations (3.15) and (3.21) in [18] seem to be incorrect, as they have no meaning.

In this paper, using the Mönch fixed point theorem, we investigate the existence of mild solutions of IVP (1.1)-(1.2). Some restricted conditions on a priori estimation and measure of noncompactness estimation have been deleted, and compactness conditions of a resolvent operator or compactness conditions on a nonlinear term f(t, X r , X r , X r ) have been weakened. Our results extend and improve some corresponding results in papers [14, 621].

2 Preliminaries

We will make the following assumptions:

( H 1 ) A generates a strongly continuous semigroup in the Banach space X.

( H 2 ) F(t)B(X), 0ta. F(t):YY and for x() continuous in Y, AF()x() L 1 [J,X]. For xX, F (t)x is continuous in tJ, where B(X) is the space of all linear and bounded operators on X, and Y is the Banach space formed from D(A), the domain of A, endowed with the graph norm.

Definition 2.1 [5]

R(t) is a resolvent operator of (1.1) with f0 if R(t)B(X) for 0ta and satisfies the following conditions:

  1. (1)

    R(0)=I, the identity operator on X,

  2. (2)

    for all xX, R(t)x is continuous for 0ta,

  3. (3)

    R(t)B(Y), 0ta; for yY, R()y C 1 [J,X]C[J,Y] and

    d d t R ( t ) y = A [ R ( t ) y + 0 t F ( t s ) R ( s ) y d s ] = R ( t ) A y + 0 t R ( t s ) A F ( s ) y d s , 0 t a .
    (2.1)

The resolvent operator R(t) is said to be equicontinuous if {tR(t)x:xB} is equicontinuous for the entire bounded set BX and t>0. If xC[J,X] satisfies the following integral equation:

x(t)=R(t) ( x 0 + g ( x ) ) + 0 t R(ts)f ( s , x ( σ 1 ( s ) ) , ( K x σ 2 ) ( s ) , ( H x σ 3 ) ( s ) ) ds,tJ,

then x is said to be a mild solution IVP (1.1)-(1.2).

Lemma 2.2 [14]

Let the conditions ( H 1 ), ( H 2 ) be satisfied. Then (1.1) with f0 has a unique resolvent operator.

The following lemma is obvious.

Lemma 2.3 Let the resolvent operator R(t) be equicontinuous. If there is ρL[J, R + ] such that x(t)ρ(t) for a.e. tJ, then the set { 0 t R(ts)x(s)ds} is equicontinuous.

Lemma 2.4 [22]

Let VC[J,E] be an equicontinuous bounded subset. Then α(V(t))C[J, R + ] ( R + =[0,)), α(V)= max t J α(V(t)).

Lemma 2.5 [23]

Let V={ x n }L[J,E] and there exists σL[J, R + ] such that x n (t)σ(t) for any xV and a.e. tJ. Then α(V(t))L[J, R + ] and

α ( { 0 t x n ( s ) d s : n N } ) 2 0 t α ( V ( s ) ) ds,tJ.

Lemma 2.6 [24] (Mönch)

Let E be a Banach space, Ω a closed convex subset in E and y 0 Ω. Suppose that the continuous operator F:ΩΩ has the following property:

VΩcountable,V co ¯ ( { y 0 } F ( V ) ) Vis relatively compact.

Then F has a fixed point in Ω.

For VC[J,X], let V(t)={x(t):xV}, V σ i (t)={x( σ i (t)):xV} (i=1,2,3), (KV)(t)={(Kx)(t):xV}, (HV)(t)={(Hx)(t):xV} (tJ), X r ={xX:xr} and S r ={xC[J,X]: x C r} for any r>0. α() and denote the Kuratowski measure of noncompactness in X and C[J,X] respectively. For details on the properties of noncompact measure, we refer the reader to [22].

3 Existence of a mild solution

We make the following assumptions for convenience.

( H 3 ) There exist constants l g >0, M>0 and 4 l g M<1 such that

g ( x ) g ( y ) l g x y C ,x,yC[J,X],

and g(0)=0.

( H 3 ) g:C[J,X]E is continuous, compact and there exists a constant N0 such that g(x)N.

( H 4 ) There exists qC[J, R + ] such that

f ( t , x , y , z ) q(t) ( x + y + z ) ,tJ,x,y,zX.

( H 5 ) There exist k 0 C[, R + ], h 0 C[J×J, R + ] such that

k ( t , s , x ) k 0 ( t , s ) x , ( t , s ) , x X , h ( t , s , x ) h 0 ( t , s ) x , t , s J , x X .

( H 6 ) For any r>0 and a bounded set V i X r , there exist constants l i >0 (i=1,2,3) such that

α ( f ( t , V 1 , V 2 , V 3 ) ) l 1 α( V 1 )+ l 2 α( V 2 )+ l 3 α( V 3 ),tJ.

( H 7 ) For any r>0 and a bounded set V X r ,

α ( k ( t , s , V ) ) k 0 ( t , s ) α ( V ) , ( t , s ) , α ( h ( t , s , V ) ) h 0 ( t , s ) α ( V ) , t , s J .

( H 8 ) The resolvent operator R(t) is equicontinuous and R(t)M e w t for tJ and some positive number

w=max { 2 M q 0 ( 1 + K 0 a + H 0 a ) , 4 M ( l 1 + 2 l 2 a K 0 + 2 l 3 a H 0 ) } ,

where K 0 = max ( t , s ) k 0 (t,s), H 0 = max t , s J h 0 (t,s), q 0 = max t J q(t).

Without loss of generality, we always suppose that x 0 =0.

Theorem 3.1 Let conditions ( H 1 ), ( H 2 ), ( H 3 )-( H 8 ) be satisfied. Then IVP (1.1)-(1.2) has at least one mild solution.

Proof Let

( F x ) ( t ) = R ( t ) g ( x ) + 0 t R ( t s ) f ( s , x ( σ 1 ( s ) ) , ( K x σ 2 ) ( s ) , ( H x σ 3 ) ( s ) ) d s , t J .
(3.1)

We have by ( H 3 ), ( H 4 ) and ( H 5 ),

(3.2)

Let

B R = { x C [ J , X ] : x C R } .

Then B R is a closed convex subset in C[J,X], 0 B R and F: B R B R . Similar to the proof of [6] and [9], it is easy to verify that F is a continuous operator from B R into B R . For x B R , sJ, ( H 4 ) and ( H 5 ) imply

f ( s , x ( σ 1 ( s ) ) , ( K x σ 2 ) ( s ) , ( H x σ 3 ) ( s ) ) q(s) ( 1 + 0 s k 0 ( s , r ) d r + 0 a h 0 ( s , r ) d r ) R.
(3.3)

We can show from (3.3), ( H 8 ) and Lemma 2.3 that F( B R ) is an equicontinuous subset in C[J,X].

Let V B R be a countable set and V co ¯ ({0}F(V)), then

V(t) co ¯ ( { 0 } ( F V ) ( t ) ) .
(3.4)

From equicontinuity of F( B R ) and (3.4), we know that V is an equicontinuous subset in C[J,X]. By the properties of noncompact measure, the conditions ( H 3 ), ( H 6 ), ( H 7 ), (3.4) and Lemma 2.5, we have

(3.5)

(3.5) together with Lemma 2.4 imply that , and so . Hence V is relatively compact in C[J,X]. Lemma 2.6 implies that F has a fixed point in C[J,X]. Then IVP (1.1)-(1.2) has at least one mild solution. The proof is completed. □

Theorem 3.2 Let the conditions ( H 1 ), ( H 2 ) and ( H 3 )-( H 8 ) be satisfied. Then IVP (1.1)-(1.2) has at least one mild solution.

Proof Similar to (3.2) and (3.5), it is easy to verify

( F x ) ( t ) MN+M q 0 (1+ K 0 a+ H 0 a) w 1 x C =MN+η x C ,

where η=M q 0 (1+ K 0 a+ H 0 a) w 1 <1. Taking R>MN ( 1 η ) 1 , let B R ={xC[J,X]: x C R}. We have F: B R B R and the inequality (3.5) is transformed into , tJ.

The other proof is similar to the proof of Theorem 3.1, we omit it. □

4 An example

Let X= L 2 [0,π]. Consider the following partial functional integro-differential equation with a nonlocal condition,

{ u t ( t , y ) = u y ( t , y ) + 0 t F ( t s ) u y ( s , y ) d s + γ 1 sin u ( t r , y ) u t ( t , y ) = + 0 t γ 2 u ( s r , y ) d s ( 1 + t ) + 0 a γ 3 u ( s r , y ) d s ( 1 + t ) ( 1 + s ) 2 , 0 t a , u ( 0 , y ) = u 0 ( y ) + γ 4 u ( y ) ,
(4.1)

where r, γ i R (i=1,2,3,4), σ 1 (t)= σ 2 (t)= σ 3 (t)=tr, 0rta, F(t) satisfies the condition ( H 2 ),

(4.2)
(4.3)
(4.4)

Let the operator A be defined by Aw= w , wD(A) with the domain

D(A)= { w E : w E , w  is almost everywhere bounded } .

Then A generates a translation semigroup R(t) and R(t) is equicontinuous. The problem (4.1) can be regarded as a form of IVP (1.1)-(1.2). We have by (4.2), (4.3) and (4.4),

f ( t , u , v , z ) | γ | ( u + v + z ) , | γ | = max { | γ 1 | , | γ 2 | , | γ 3 | } , u , v , z X , k ( t , s , u ) u , h ( t , s , u ) u , u X ,

and

g ( u ) g ( v ) | γ 4 | u v C ,g(0)=0.

γ 4 and M can be chosen such that 4M| γ 4 |<1. In addition, for any r>0 and a bounded set V i X r (i=1,2,3), we can show that by the diagonal method,

α ( f ( t , V 1 , V 2 , V 3 ) ) | γ | ( α ( V 1 ) + α ( V 2 ) + α ( V 3 ) ) , t J , α ( k ( t , s , V 1 ) ) α ( V 1 ) , t , s , α ( h ( t , s , V 1 ) ) α ( V 1 ) , t , s [ 0 , a ] .

Hence all the conditions of Theorem 3.1 are satisfied, the problem (4.1) has at least one mild solution in C[J,X].