1 Introduction and preliminaries

Throughout this article we assume that E is a real Banach space with norm ||·||, E* is the dual space of E; 〈·, ·〉 is the duality pairing between E and E*; C is a nonempty closed convex subset of E; ℕ denotes the natural number set; ℜ+ is the set of nonnegative real numbers; The mapping J:E 2 E * defined by

J ( x ) = f * E * : x , f * = x 2 ; f * = x , x E
(1)

is called the normalized duality mapping. We denote a single valued normalized duality mapping by j.

Let T: CC be a nonlinear mapping; F(T) denotes the set of fixed points of mapping T, i.e., F(T) := {xC, x = Tx}. We use "→" to stand for strong convergence and "⇀" for weak convergence. For a given sequence {x n } ⊂ C, let ω w (x n ) denote the weak ω-limit set.

Recall that T is said to be pseudocontractive if for all x, yC, there exists j(x - y) ∈ J(x - y) such that

T x - T y , j ( x - y ) x - y 2 ;
(2)

T is said to be strongly pseudocontr active if there exists a constant α ∈ (0,1), such that for any x, yC, there exists j(x - y) ∈ J(x - y)

T x - T y , j ( x - y ) α x - y 2 .
(3)

In recent years, many authors have focused on the studies about the existence and convergence of fixed points for the class of pseudocontractions. Especially in 1974, Deimling [1] proved the following existence theorem of fixed point for a continuous and strong pseudocontraction in a nonempty closed convex subset of Banach spaces.

Theorem D. Let E be a Banach space, C be a nonempty closed convex subset of E and T: CC be a continuous and strong pseudocontraction. Then T has a unique fixed point in C.

Recently, the problems of convergence of an implicit iterative algorithm to a common fixed point for a family of nonexpansive mappings or pseudocontractive mappings have been considered by several authors, see [25]. In 2001, Xu and Ori [2] firstly introduced an implicit iterative x n = α n xn-1+ (1 - α n )T n x n , n ∈ ℕ, x0C for a finite family of nonexpansive mappings T i i = 1 N and proved some weak convergence theorems to a common fixed point for a finite family of nonexpansive mappings in a Hilbert space. In 2004, Osilike [3] improved the results of Xu and Ori [2] from nonexpansive mappings to strict pseudocontractions in the framework of Hilbert spaces. In 2006, Chen et al. [4] extended the results of Osilike [3] to more general Banach spaces.

On the other hand, the convergence problems of semi-groups have been considered by many authors recently. Suzuki [6] considered the strong convergence to common fixed points of nonexpansive semigroups in Hilbert spaces. Xu [7] gave strong convergence theorem for contraction semigroups in Banach spaces. Chang et al. [8] proved the strong convergence theorem for nonexpansive semi-groups in Banach space. He also studied the weak convergence problems of the implicit iteration process for Lipschitzian pseudocontractive semi-groups in the general Banach spaces [9]. The pseudocontractive semi-groups is defined as follows.

Definition 1.1 (1) One-parameter family T: = {T(t): t ≥ 0} of mappings from C into itself is said to be a pseudo-contraction semigroup on C, if the following conditions are satisfied:

(a). T(0)x = x for each xC;

(b). T(t + s)x = T(s)T(t) for any t, s ∈ ℜ+ and xC;

(c). For any xC, the mapping tT(t)x is continuous;

(d). For all x, yC, there exists j(x - y) ∈ J(x - y) such that

T ( t ) x - T ( t ) y , j ( x - y ) x - y 2 , f o r a n y t > 0 .
(4)

(2) A pseudo-contraction semigroup of mappings from C into itself is said to be a Lipschitzian if the condition (a)-(d) and following condition (f) are satisfied.

(f) there exists a bounded measurable function L: [0, ∞) → [0, ∞) such that for any x, yC,

T ( t ) x - T ( t ) y L ( t ) x - y

for any t > 0. In the sequel, we denote it by

L = sup t 0 L ( t ) <
(5)

Cho et al. [10] considered viscosity approximations with continuous strong pseudocontractions for a pseudocontraction semigroup and prove the following theorem.

Theorem Cho. Let E be a real uniformly convex Banach space with a uniformly Gâ teaux differentiable norm, and C be a nonempty closed convex subset of E. Let T(t): t ≥ 0 be a strongly continuous L-Lipschitz semigroup of pseudocontractions on C such that Ω , where Ω is the set of common fixed points of semi-group T(t). Let f: CC be a fixed bounded, continuous and strong pseudocontraction with the coefficient α in (0,1), let α n and t n be sequences of real numbers satisfying α n ∈ (0, 1), t n > 0, and lim n t n = lim n α n t n =0; Let {x n } be a sequence generated in the following manner:

x n = ( 1 - α n ) f ( x n ) + α n T ( t n ) x n , n 1 .
(6)

Assume that LIM||T(t)x n - T(t)x*|| ≤ ||x n - x*||, ∀x*K, t ≥ 0, where K := {x*C: Φ(x*) = minxCΦ(x)} with Φ(x) = LIM||x n - x||2, ∀xC. Then x n converges strongly to x* ∈ Ω which solves the following variational inequality: 〈(I - f)x*, j(x* - x)〉 ≤ 0, ∀x ∈ Ω.

Qin and Cho [11] established the theorems of weak convergence of an implicit iterative algorithm with errors for strongly continuous semigroups of Lipschitz pseudocontractions in the framework of real Banach spaces.

Theorem Q. Let E be a reflexive Banach space which satisfies Opial's condition and K a nonempty closed convex subset of E. Let T:= { T ( t ) : t 0 } be a strongly continuous semigroup of Lipschitz pseudocontractions from K into itself with F:= t 0 F ( T ( t ) ) ; Assume that supt≥0{L(t)} < ∞, where L(t) is the Lipschitz constant of the mapping T(t). Let {x n } be a sequence generated by the following iterative process:

x 0 K ; x n = α n x n - 1 + β n T ( t n ) x n + γ n u n ; n 1 ;
(7)

where {α n }, {β n }, {γ n } are sequences in (0,1), {t n } is a sequence in (0, ∞) and {u n } is a bounded sequence in K. Assume that the following conditions are satisfied:

(a) α n + β n + γ n = 1;

(b) lim n t n = lim n α n + γ n t n =0.

Then the sequence {x n } generated in (7) converges weakly to a common fixed point of the semigroup T:= { T ( t ) : t 0 } ;

Agarwal et al. [12] studied strongly continuous semigroups of Lipschitz pseudocontractions and proved the strong convergence theorems of fixed points in an arbitrary Banach space based on an implicit iterative algorithm.

Theorem A. Let E be an arbitrary Banach space and K a nonempty closed convex subset of E. Let T:= { T ( t ) : t 0 } be a strongly continuous semigroup of Lipschitz pseudocontractions from K into itself with F:= t 0 F ( T ( t ) ) . Assume that supt≥0{L(t)} < ∞, where L(t) is the Lipschitz constant of the mapping T(t). Let {x n } be a sequence in

x 0 K ; x n = α n x n - 1 + β n T ( t n ) x n + γ n u n ; n 1 ,
(8)

where {α n }, {β n }, {γ n } are sequences in (0,1) such that α n + β n + γ n = 1, {t n } is a sequence in (0, ∞) and {u n } is a bounded sequence in K. Assume that lim n t n = lim n α n + γ n t n =0, lim n γ n α n + γ n < and there is a nondecreasing function f: (0, ∞) → (0, ∞) with f(0) = 0 and f(t) > 0 for all t ∈ (0, ∞) such that, for all xC, sup { x - T ( t ) x : t 0 } f ( dist ( x , F ) ) . Then the sequence {x n } converges strongly to a common fixed point of the semigroup T:= { T ( t ) : t 0 } .

The purpose of this article is to prove the strong and weak convergence of implicit iterative process

x n = ( 1 - α n ) x n - 1 + α n T ( t n ) x n , n , x 0 C
(9)

for a pseudocontraction semigroup T: = {T(t): t ≥ 0} in the framework of Banach spaces, which improves and extends the corresponding results of many author's. We need the following Lemma.

Lemma 1.1 [9] Let E be a real reflexive Banach space with Opial condition. Let C be a nonempty closed convex subset of E and T: CC be a continuous pseudocontractive mapping. Then I - T is demiclosed at zero, i.e., for any sequence {x n } ⊂ E, if x n y and ||(I - T)x n || → 0, then (I - T)y = 0.

2 Main results

Theorem 2.1 Let E be a real Banach space and C be a nonempty compact convex subset of E. Let T: = {T(t): t ≥ 0}: CC be a Lipschitian and pseudocontraction semigroup defined by Definition 1.1 with a bounded measurable function L: [0, ∞) → [0, ∞). Suppose F ( T ) := t 0 F ( T ( t ) ) . Let α n and t n be sequences of real numbers satisfying t n > 0, α n ∈ [a, 1) ⊂ (0, 1) and limn→∞α n = 1. Then the sequence {x n } defined by (9) converges strongly to a common fixed point x*F(T) in C.

Proof. We divide the proof into five steps.

(I). The sequence {x n } defined by x n = (1 - α n )xn-1+ α n T(t n )x n , n ∈ ℕ, x0C is well defined.

In fact for all n ∈ ℕ, we define a mapping S n as follows:

S n x = ( 1 - α n ) x n - 1 + α n T ( t n ) x , n , x C .
(10)

Then we have

S n x - S n y , j ( x - y ) = α n T ( t n ) x - T ( t n ) y , j ( x - y ) α n x - y 2 .
(11)

So S n is strongly pseudo-contraction, thus from Theorem D, there exists a point x n such that x n = (1 - α n )xn-1+ α n T(t n )x n , that is the sequence {x n } defined by x n = (1 - α n )xn-1+ α n T(t n )x n , n ∈ ℕ, x0C is well defined.

(II). Since the common fixed-point set F(T) is nonempty let pF(T). For each pF(T), we prove that limn→∞||x n - p|| exists.

In fact

x n - p 2 = x n - p , j ( x n - p ) = ( 1 - α n ) ( x n - 1 - p ) + α n ( T ( t n ) x n - p ) , j ( x - p ) ( 1 - α n ) x n - 1 - p x n - p + α n x n - p 2 .
(12)

So we get ||x n - p|| ≤ (1 - α n )||xn-1- p|| + α n ||x n - p||, that is

x n - p x n - 1 - p .

This implies that the limit limn→∞||x n - p|| exists.

(III). We prove limn→∞||T(t n )x n - x n || = 0.

The sequence {||x n - p||n∈ℕ} is bounded since limn→∞||x n - p|| exists, so the sequence {x n } is bounded. Since

T ( t n ) x n = x n - ( 1 - α n ) x n - 1 α n x n α n + ( 1 - α n ) x n - 1 α n x n a + ( 1 - α n ) x n - 1 a ,
(13)

This shows that {T(t n )x n } is bounded. In view of

x n - T ( t n ) x n = ( 1 - α n ) ( x n - 1 - T ( t n ) x n ) = 1 - α n x n - 1 - T ( t n ) x n

and condition limn→∞α n = 1, we have

lim n T ( t n ) x n - x n = 0 .
(14)

(IV). Now we prove that for all t > 0, limn→∞||T(t)x n - x n || = 0.

Since pseudocontraction semigroup T: = {T(t) : t ≥ 0} is Lipschitian, for any k ∈ ℕ,

T ( ( k + 1 ) t n ) x n - T ( k t n ) x n = T ( k t n ) T ( t n ) x n - T ( k t n ) x n L ( k t n ) T ( t n ) x n - x n L T ( t n ) x n - x n .
(15)

Because limn→∞||T(t n )x n - x n || = 0, so for any k ∈ ℕ,

lim n T ( ( k + 1 ) t n ) x n - T ( k t n ) x n = 0 .
(16)

Since

T ( t ) x n - T t t n t n x n = T t t n t n T t - t t n t n x n - T t t n t n x n L T t - t t n t n x n - x n
(17)

and T(·) is continuous, we have

lim n T t t n t n x n - T ( t ) x n = 0 .
(18)

So from

x n - T ( t ) x n k = 0 t t n - 1 T ( ( k + 1 ) t n ) x n - T ( k t n ) x n + T t t n t n x n - T ( t ) x n ,
(19)

and limn→∞||T((k+1)t n )x n - T(kt n )x n || = 0 as well as lim n T t t n t n x n - T ( t ) x n =0, we can get

lim n T ( t ) x n - x n = 0 .
(20)

(V). We prove {x n } converges strongly to an element of F(T).

Since C is a compact convex subset of E, we know there exists a subsequence x n j x n , such that x n j xC. So we have lim j T ( t ) x n j - x n j =0 from limn→∞||T(t)x n - x n || = 0, and

x - T ( t ) x = lim j T ( t ) x n j - x n j = 0 .
(21)

This manifests that xF(T). Because for any pF(T), limn→∞||x n - p|| exists, and lim n x n - x = lim j x n j - x =0, we have that {x n } converges strongly to an element of F(T). This completes the proof of Theorem 2.1.

Theorem 2.2 Let E be a reflexive Banach space satisfying the Opial condition and C be a nonempty closed convex subset of E. Let T: = {T(t): t ≥ 0}: CC be a Lipschitian and pseudocontraction semigroup defined by Definition 1.1 with a bounded measurable function L: [0, ∞) → [0, ∞). Suppose F ( T ) := t 0 F ( T ( t ) ) . Let α n and t n be sequences of real numbers satisfying t n > 0, α n ∈ [a, 1) ⊂ (0,1) and limn→∞α n = 1. Then the sequence {x n } defined by x n = (1 - α n )xn-1+ α n T(t n )x n , x0C, n ∈ ℕ, converges weakly to a common fixed point x*F(T) in C.

Proof. It can be proved as in Theorem 2.1, that for each pF(T), the limit limn→∞||x n - p|| exists and {T(t n )x n } is bounded, for all t > 0, limn→∞||T(t)x n - x n || = 0. Since E is reflexive, C is closed and convex, {x n } is bounded, there exist a subsequence x n j x n such that x n j x. For any t > 0, we have lim n j T ( t ) x n j - x n j =0. By Lemma 1.1, xF(T(t)), ∀t > 0. Since the space E satisfies Opial condition, we see that ω w (x n ) is a singleton. This completes the proof.

Remark 2.1 There is no other condition imposed on t n in the Theorems 2.1 and 2.2 except that in the definition of pseudo-contraction semigroups. So our results improve corresponding results of many authors such as [1012], of cause extend many results in [48].