Skip to main content

Nonlinear ergodic theorems and weak convergence theorems for reversible semigroup of asymptotically nonexpansive mappings in Banach spaces

Abstract

In this paper, we provide the nonlinear ergodic theorems and weak convergence theorems for almost orbits of a reversible semigroup of asymptotically nonexpansive mappings in a uniformly convex Banach space without assuming that X has a Fréchet differentiable norm. Since almost orbits in this paper are not almost asymptotically isometric, new methods have to be introduced and used for the proofs. Our main results include many well-known results as special cases and are new even for reversible semigroup of nonexpansive mappings.

MSC:47H20.

1 Introduction

Baillon [1] proved the first nonlinear ergodic theorem for nonexpansive mappings in the framework of Hilbert space. Baillon′s theorem was extended to various semigroups in Hilbert spaces [24] or Banach spaces [513]. For instance, Takahashi [2] proved the ergodic theorem for right reversible semigroups of nonexpansive mappings in a Hilbert space by using the methods of invariant means. Lau et al. [5] studied the existence of nonexpansive retractions for amenable semigroups of nonexpansive mappings and provided the nonlinear ergodic theorems in Banach spaces. Kim and Li [6] proved the ergodic theorem for the almost asymptotically isometric almost orbits of right reversible semigroups of asymptotically nonexpansive mappings in a uniformly convex Banach space with a Fréchet differentiable norm. Many papers about weak convergence of asymptotically nonexpansive semigroups in a uniformly convex Banach space with a Fréchet differentiable norm have appeared [6, 10, 11, 1416]. In 2001, Falset, et al. [14], Kaczor [15] proved the weak convergence theorems of almost orbits of commutative semigroups of asymptotically nonexpansive mappings under the assumptions that the Banach space is uniformly convex, and its dual space has the Kadec-Klee property.

This paper is devoted to the study of the nonlinear ergodic theorem and weak convergence for almost orbits of reversible semigroups of asymptotically nonexpansive mappings. Using the technique of product net, we first obtain the nonlinear ergodic theorems without assuming that the uniformly convex Banach space has a Fréchet differentiable norm, which extend and unify many previously known results in [2, 6, 10, 11, 16]. Next, we establish the convergence theorem in the case of reversible semigroup and the uniformly convex Banach space whose dual space has the Kadec-Klee property, which improves the known ones (see [2, 11, 1416]) for commutative semigroups of asymptotically nonexpansive mappings in a uniformly convex Banach space. It is safe to say that the many general and key assumptions in the situation of reversible semigroup, such as the almost orbit u() is almost asymptotically isometric, and the subspace D has a left invariant mean (see [2, 6, 10]), are not necessary in this paper. Our main results are new even for the reversible semigroup of nonexpansive mappings.

2 Preliminaries

Let C be a nonempty bounded closed convex subset of a Banach space X. Let X be the dual of X, then the value of x X at xX will be denoted by x, x , and we associate the set

J(x)= { x X : x , x = x 2 = x 2 } .

It is clear from the Hahn-Banach theorem that J(x) for all xX. Then the multi-valued operator J:X X is called the normalized duality mapping of X. We say that X has a Fréchet differentiable norm, i.e., for each x0, lim t 0 (x+tyx)/t exists uniformly in y B r ={zX:zr}, r>0. We say that X has the Kadec-Klee property if for every sequence { x n } n N in X, whenever ω- lim n x n =x with lim n x n =x, it follows that lim n x n =x. Recall that X has the Kadec property if for every net { x α } α I in X, whenever ω- lim α I x α =x with lim α I x α =x, it follows that lim α I x α =x, where I is a directed system. It is well known that within the class of reflexive spaces, the Kadec-Klee property is equivalent to the Kadec property [17]. We also would like to remark that a uniformly convex Banach space with a Fréchet differentiable norm implies that its dual has Kadec-Klee property, while the converse implication fails [14, 15].

Let G be a semitopological semigroup, i.e., G is a semigroup with a Hausdorff topology such that for each tG, the mappings sst and sts from G to G are continuous. G is called right reversible if any two closed left ideals of G have nonvoid intersection. In this case, (G,) is a directed system when the binary relation ≤ on G is defined by st if and only if {s} G s ¯ {t} G t ¯ , s,tG. Right reversible semitopological semigroups include all commutative semigroups and all semitopological semigroups, which are right amenable as discrete semigroups.

Let m(G) be the Banach space of all bounded real valued functions on G with the supremum norm. Then for each sG and fm(G), we can define l s f in m(G) by ( l s f)(t)=f(st) for all tG. Let D be a subspace of m(G) containing constant functions and invariant under l s for every sG. Let D be the dual space of D, then the value of μ D at fD will be denoted by μ(f)=f(t)dμ(t)=μ(t)f(t). A linear function μ on D is called a mean on D if μ=μ(1)=1. Further, a mean μ on D is left invariant if for all sG and fD, μ( l s f)=μ(f). For each sG, we define a point evaluation δ s on D by δ s (f)=f(s) for every fD. A convex combination of point evaluation is called a finite mean on G.

By Day [18], if D has a left invariant mean, then there exists a net { λ α :αA} of finite means on G such that

lim α A λ α l s λ α =0

for every sG, where A is a directed system, and l s is the conjugate operator of l s .

Let ={T(t):tG} be a semigroup acting on C, i.e., T(ts)x=T(t)T(s)x for all t,sG and xC. Recall that is said to be asymptotically nonexpansive [1921] if there exists a function α():G[0,+) with lim sup t G α(t)=0 such that for all x,yC and tG,

T ( t ) x T ( t ) y ( 1 + α ( t ) ) xy.

If α(t)0 for every tG, then is said to be nonexpansive. It should be pointed out that there is a notion of asymptotically nonexpansive mappings defined dependent on right ideals in a semigroup in [22, 23].

A function u():GC is said to be an almost orbit of [16] if

lim sup t G [ sup h G u ( h t ) T ( h ) u ( t ) ] =0.

Suppose that u() is an almost orbit of such that for each x X , the function h x :tu(t), x is in D. For each μ D , since X is reflexive, there exists a unique u μ in X such that u μ , x =u(t), x dμ(t) for all x X . We denote u μ by μ(t)u(t). If λ is a finite mean on G, say λ= i = 1 n a i δ s i , where s i G, a i 0, i=1,2,,n, and i = 1 n a i =1, then

λ(t) u ( t ) = i = 1 n a i u( s i ).

Throughout this paper, let C be a nonempty bounded closed convex subset of uniformly convex Banach space X, and let ={T(t):tG} be a reversible semigroup of asymptotically nonexpansive mappings acting on C. Let F() denote the set of all fixed points of , i.e., F()={xC:T(t)x=x for all tG}. For each ε>0 and hG, we set

F ε ( T ( h ) ) = { x C : T ( h ) x x ε } .

It should be noted that if for any ε>0, there exists h ε G such that for all h h ε , x F ε (T(h)), then lim h G T(h)x=x, and thus xF() by the continuity of {T(h),hG}.

We denote by AO() the set of all almost orbits of and by LAO() the set {T(h)u():hG,uAO()}. Denote by ω ω (u) the set of all weak limit points of subnets of the net { u ( t ) } t G .

3 Lemmas

In this section, we prove some lemmas, which play a crucial role in the proof of our main theorems in next section.

Lemma 3.1 [24]

Let X be a Banach space and J be the normalized duality mapping. Then for given x,yX, the following inequality holds:

x + y 2 x 2 +2 y , j ( x + y )

for all j(x+y)J(x+y).

Lemma 3.2 [25]

Let C be a nonempty bounded closed convex subset of a uniformly convex Banach space X. Then there exists a strictly increasing continuous convex function γ:[0,+)[0,+) with γ(0)=0 such that

γ ( T ( i = 1 n a i x i ) i = 1 n a i T x i ) max 1 i , j n { x i x j T x i T x j }

for all integers n1, a 1 ,, a n 0 with i = 1 n a i =1, x 1 ,, x n C, and all nonexpansive mapping T:CC.

From Lemma 3.2, we can get for all a 1 ,, a n 0 with i = 1 n a i =1, x 1 ,, x n C,

T ( h ) ( i = 1 n a i x i ) i = 1 n a i T ( h ) x i ( 1 + α ( h ) ) γ 1 ( max 1 i , j n { x i x j 1 1 + α ( h ) T ( h ) x i T ( h ) x j } ) ( 1 + α ( h ) ) γ 1 ( max 1 i , j n { x i x j T ( h ) x i T ( h ) x j } + d α ( h ) ) ,

where d=4sup{x:xC}+1.

To simplify, in the following, for each ε(0,1], we define

a(ε)=min { ε 2 ( d + 2 ) 2 , ε 3 ( 3 d + 2 ) 2 γ ( ε 4 ) }

and

G ε = { h G : α ( h ) ε } ,

where γ() is as in Lemma 3.2. Then G ε is nonempty for each ε>0, and if h G ε , then for all th, t G ε . And it should also be noted G a ( ε ) G ε for all ε(0,1].

Lemma 3.3 For all h G a ( ε ) ,

co ¯ F a ( ε ) ( T ( h ) ) F ε ( T ( h ) ) .

Proof Since F ε (T(h)) is closed, we only need to prove that for all h G a ( ε ) ,

co F a ( ε ) ( T ( h ) ) F ε ( T ( h ) ) .

Let y= i = 1 n a i y i , y i F a ( ε ) (T(h)), a i 0, i=1,,n, and i = 1 n a i =1. Then

T ( h ) y y = T ( h ) i = 1 n a i y i i = 1 n a i y i T ( h ) i = 1 n a i y i i = 1 n a i T ( h ) y i + i = 1 n a i T ( h ) y i i = 1 n a i y i 2 γ 1 ( max 1 i , j n { y i y j T ( h ) y i T ( h ) y j } + d α ( h ) ) + a ( ε ) 2 γ 1 ( max 1 i , j n { y i T ( h ) y i + y j T ( h ) y j } + d α ( h ) ) + a ( ε ) 2 γ 1 ( 2 a ( ε ) + d a ( ε ) ) + a ( ε ) ε 2 + ε 2 = ε .

This completes the proof. □

Lemma 3.4 For all h G ε 4 ,

F ε 4 ( T ( h ) ) +B ( 0 , ε 4 ) F ε ( T ( h ) ) .

Proof Let h G ε 4 and x=y+z F ε 4 (T(h))+B(0, ε 4 ), where y F ε 4 (T(h)) and zB(0, ε 4 ), then

T ( h ) x x = T ( h ) ( y + z ) ( y + z ) T ( h ) ( y + z ) T ( h ) y + T ( h ) y y + z 2 z + T ( h ) y y + z 3 ε 4 + ε 4 = ε .

This completes the proof. □

Lemma 3.5 Let ε(0,1] and h G a ( a ( ε 4 ) ) , then there exists an n 0 N such that for all n n 0 and xC,

1 n i = 1 n T ( h i ) x F ε ( T ( h ) ) .

Proof Let ε(0,1] and m= 2 d + 1 a ( ε 4 ) , there is an n 0 N satisfying

n 0 max { 12 m d ε , 32 m 2 d ( d + 1 ) ( γ ( a ( ε 4 ) 2 ) ε ) 1 } .

For any given n n 0 and h G a ( a ( ε 4 ) ) , we can take a number

K= m 2 d ( 1 + 2 n α ( h ) ) ( γ ( a ( ε 4 ) 2 ) ) 1 ( K < n 2 ) .

For each iN and xC, we set

a i (x)=γ ( 8 9 1 m j = 1 m T ( h i + j + 1 ) x T ( h ) 1 m j = 1 m T ( h i + j ) x ) .

Noting α(h) 1 8 and

a i ( x ) max 1 j , k m { T ( h i + j ) x T ( h i + k ) x T ( h i + j + 1 ) x T ( h i + k + 1 ) x + d α ( h ) } 1 j < k m ( T ( h i + j ) x T ( h i + k ) x T ( h i + j + 1 ) x T ( h i + k + 1 ) x + d α ( h ) ) ,

we get

i = 1 n a i ( x ) i = 1 n 1 j < k m ( T ( h i + j ) x T ( h i + k ) x T ( h i + j + 1 ) x T ( h i + k + 1 ) x + d α ( h ) ) = 1 j < k m i = 1 n ( T ( h i + j ) x T ( h i + k ) x T ( h i + j + 1 ) x T ( h i + k + 1 ) x + d α ( h ) ) 1 j < k m ( d + n d α ( h ) ) m 2 d ( 1 + n α ( h ) ) .

Suppose that there are k elements in { a i (x):i=1,2,,2n} such that a i (x)γ( a ( ε 4 ) 2 ), then

kγ ( a ( ε 4 ) 2 ) m 2 d ( 1 + 2 n α ( h ) ) .

Hence

k m 2 d ( 1 + 2 n α ( h ) ) ( γ ( a ( ε 4 ) 2 ) ) 1 =K.

Thus, there are at most N=[K] terms in { a i (x):i=1,2,,2n} with a i (x)γ( a ( ε 4 ) 2 ). Therefore, for each i{1,2,,n}, there is at least one term a i + j 0 (x) (0 j 0 N) in { a i + j (x):j=0,1,,N} satisfying a i + j 0 (x)<γ( a ( ε 4 ) 2 ).

Putting

l i =min { j : a i + j ( x ) < γ ( a ( ε 4 ) 2 ) , 0 j N } ,

i=1,2,,n. It is easy to see that there are at most N elements in {i:i=1,2,,n} such that l i 0. Since

T ( h ) 1 m j = 1 m T ( h i + l i + j ) x 1 m j = 1 m T ( h i + l i + j ) x T ( h ) 1 m j = 1 m T ( h i + l i + j ) x 1 m j = 1 m T ( h i + l i + j + 1 ) x + 1 m j = 1 m T ( h i + l i + j ) x 1 m j = 1 m T ( h i + l i + j + 1 ) x 9 8 γ 1 ( a i + l i ( x ) ) + d 2 m 9 16 a ( ε 4 ) + 1 4 a ( ε 4 ) < a ( ε 4 ) ,

we can conclude that for all h G a ( a ( ε 4 ) ) ,

1 m j = 1 m T ( h i + l i + j ) x F a ( ε 4 ) ( T ( h ) ) .

By Lemma 3.3, we get for all h G a ( a ( ε 4 ) ) G a ( ε 4 ) ,

1 n i = 1 n 1 m j = 1 m T ( h i + l i + j ) xco F a ( ε 4 ) ( T ( h ) ) F ε 4 ( T ( h ) ) .

Using Lemma 3.4 and

1 n i = 1 n T ( h i ) x 1 n i = 1 n 1 m j = 1 m T ( h i + l i + j ) x 1 m n j = 1 m i = 1 n T ( h i ) x i = 1 n T ( h i + l i + j ) x 1 m n j = 1 m i = 1 n T ( h i ) x i = 1 n T ( h i + j ) x + 1 m n j = 1 m i = 1 n T ( h i + j ) x i = 1 n T ( h i + l i + j ) x m d n + N d n ε 12 + m 2 d 2 ( γ ( a ( ε 4 ) 2 ) ) 1 n + 2 m 2 d 2 α ( h ) ( γ ( a ( ε 4 ) 2 ) ) 1 < ε 12 + ε 32 + ε 8 < ε 4 ,

we obtain

1 n i = 1 n T ( h i ) x F ε 4 ( T ( h ) ) +B ( 0 , ε 4 ) F ε ( T ( h ) ) .

This completes the proof. □

Lemma 3.6 Let u() be an almost orbits of . Then

lim t G λ u ( t ) + ( 1 λ ) f g

exists for all λ(0,1) and f,gF().

Proof We only need to show that

inf s G sup t G λ u ( t s ) + ( 1 λ ) f g sup s G inf t G λ u ( t s ) + ( 1 λ ) f g .

In fact, for any ε>0, there are t 0 and s 0 G such that for any tG, α(t t 0 )< ε 1 + d and φ(t s 0 )<ε, where φ(t)= sup h G u(ht)T(h)u(t). Then for all aG,

inf s G sup t G u ( t s s 0 ) f sup t G u ( t t 0 a s 0 ) f sup t G u ( t t 0 a s 0 ) T ( t t 0 ) u ( a s 0 ) + sup t G T ( t t 0 ) u ( a s 0 ) f φ ( a s 0 ) + sup t G ( 1 + α ( t t 0 ) ) u ( a s 0 ) f u ( a s 0 ) f + 2 ε .

Hence inf s G sup t G u(ts s 0 )f inf a G u(a s 0 )f+2ε. Thus, there exists s 1 G such that

sup t G u ( t s 1 s 0 ) f < inf a G u ( a s 0 ) f +3ε.

So, for any aG, we get

inf s G sup t G λ u ( t s ) + ( 1 λ ) f g sup t G λ u ( t t 0 a s 1 s 0 ) + ( 1 λ ) f g λ sup t G u ( t t 0 a s 1 s 0 ) T ( t t 0 ) u ( a s 1 s 0 ) + sup t G λ T ( t t 0 ) u ( a s 1 s 0 ) + ( 1 λ ) f g φ ( a s 1 s 0 ) + sup t G λ T ( t t 0 ) u ( a s 1 s 0 ) + ( 1 λ ) f T ( t t 0 ) ( λ u ( a s 1 s 0 ) + ( 1 λ ) f ) + sup t G T ( t t 0 ) ( λ u ( a s 1 s 0 ) + ( 1 λ ) f ) g ε + sup t G ( 1 + α ( t t 0 ) ) γ 1 ( u ( a s 1 s 0 ) f T ( t t 0 ) u ( a s 1 s 0 ) f + d α ( t t 0 ) ) + sup t G ( 1 + α ( t t 0 ) ) λ u ( a s 1 s 0 ) + ( 1 λ ) f g ε + ( 1 + ε ) sup t G γ 1 ( u ( a s 1 s 0 ) f u ( t t 0 a s 1 s 0 ) f + φ ( a s 1 s 0 ) + ε ) + ( 1 + ε ) λ u ( a s 1 s 0 ) + ( 1 λ ) f g ε + ( 1 + ε ) γ 1 ( 5 ε ) + ( 1 + ε ) λ u ( a s 1 s 0 ) + ( 1 λ ) f g .

Hence we have

inf s G sup t G λ u ( t s ) + ( 1 λ ) f g ε + ( 1 + ε ) γ 1 ( 5 ε ) + ( 1 + ε ) inf a G λ u ( a s 1 s 0 ) + ( 1 λ ) f g ε + ( 1 + ε ) γ 1 ( 5 ε ) + ( 1 + ε ) sup b G inf a G λ u ( a b ) + ( 1 λ ) f g .

Since ε>0 is arbitrary, we can conclude

inf s G sup t G λ u ( t s ) + ( 1 λ ) f g sup s G inf t G λ u ( t s ) + ( 1 λ ) f g .

This completes the proof. □

4 Main results

Theorem 4.1 Let X be a uniformly convex Banach space, and let C be a nonempty bounded closed convex subset of X. Let ={T(t):tG} be a reversible semigroup of asymptotically nonexpansive mappings on C. If D has a left invariant mean, then there exists a retraction P from LAO() onto F() satisfying the following properties:

  1. (1)

    P is nonexpansive in the sense

    PuPv inf s G sup t G u ( s t ) v ( s t ) ,u,vLAO();
  2. (2)

    PT(h)u=T(h)Pu=Pu for all uAO() and hG;

  3. (3)

    Pu s G conv ¯ {u(t):ts} for all uLAO().

Proof Since D has a left invariant mean, there exists a net { λ α :αA} of finite means on G such that lim α A λ α l s λ α =0 for every sG, where A is a directed system. Put I=A×G={β=(α,t):αA,tG}. For β i =( α i , t i )I, i=1,2, we define β 1 β 2 if and only if α 1 α 2 , t 1 t 2 . In this case, I is also a directed system. For each β=(α,t)I, define P 1 β=α, P 2 β=t and λ β = λ α . Then for every sG,

lim β I λ β l s λ β =0.
(4.1)

Let Λ={ { t β } β I , t β P 2 β,βI}. Taking any { t β ,βI}Λ, since r t β λ β is bounded, without loss of generality, suppose that r t β λ β is weakly convergent. Then for all uLAO(), ω- lim β I λ β (t)u(t t β ) exists. We define

Pu=ω- lim β I λ β (t) u ( t t β ) .

It is easy to see that for all uLAO(), Pu s G conv ¯ {u(t):ts}. Next, we shall show that PuF(). In fact, for any given ε(0,1], there exists t 0 G such that for any t t 0 , φ(t)< a ( ε ) 4 . Also, we can suppose that P 2 β t 0 for all βI, then t β t 0 , { t β }Λ. By Lemma 3.5, for any h G a ( a ( a ( ε ) 16 ) ) , there exists nN such that for all tG and βI,

1 n i = 1 n T ( h i ) u(t t β ) F a ( ε ) 4 ( T ( h ) ) .

Noting for all tG,

1 n i = 1 n T ( h i ) u ( t t β ) 1 n i = 1 n u ( h i t t β ) φ(t t β )< a ( ε ) 4 ,

we have for any h G a ( a ( a ( ε ) 16 ) ) ,

1 n i = 1 n u ( h i t t β ) F a ( ε ) 4 ( T ( h ) ) +B ( 0 , a ( ε ) 4 ) F a ( ε ) ( T ( h ) ) .

By (4.1), we obtain

lim β I λ β ( t ) 1 n i = 1 n u ( h i t t β ) λ β ( t ) u ( t t β ) =0.

Combining it with the definition of Pu, we get for all h G a ( a ( a ( ε ) 16 ) ) ,

Pu=ω- lim β I λ β (t) 1 n i = 1 n u ( h i t t β ) co ¯ F a ( ε ) ( T ( h ) ) .

Thus, by the Lemma 3.3, we can conclude that for all h G a ( a ( a ( ε ) 16 ) ) , Pu F ε (T(h)). The continuity of T(h) then implies that PuF(). Obviously, for any hG,

P T ( h ) u = ω - lim β I λ β ( t ) T ( h ) u ( t t β ) = ω - lim β I λ β ( t ) u ( h t t β ) = ω - lim β I λ β ( t ) u ( t t β ) (by (4.1)) = P u

and for any vLAO() and sG,

P u P v lim inf β I λ β ( t ) u ( t t β ) λ β ( t ) v ( t t β ) = lim inf β I λ β ( t ) u ( s t t β ) λ β ( t ) v ( s t t β ) (by (4.1)) lim inf β I λ β ( t ) sup t G u ( s t t β ) v ( s t t β ) sup t G u ( s t ) v ( s t ) .

Thus,

PuPv inf s G sup t G u ( s t ) v ( s t ) .

This completes the proof. □

Remark 4.1 It should be noted that in Theorem 4.1, we do not assume F(). In fact, we can find a fixed point PuF(). It also should be pointed out that in the case of reversible semigroup, if D has a left invariant mean, then F() (see [[6], Theorem 3.1] and [[10], Lemma 4]).

As in [6], we have the following ergodic theorem.

Theorem 4.2 Let X be a uniformly convex Banach space and C a nonempty bounded closed convex subset of X. Let ={T(t):tG} of a reversible semigroup of asymptotically nonexpansive mappings on C. If D has a left invariant mean and there exists a unique retraction P from LAO() onto F(), which satisfies the properties (1)-(3) in Theorem  4.1, then for every strongly regular net { μ α :αA} on D and uAO(),

ω- lim α A u(th)d μ α (t)=pF()umiforly in hΛ(G),

where Λ(G)={sG:st=ts for all tG}.

Remark 4.2 By Theorem 4.1 and Theorem 4.2, we can get many known theorems in [2, 6, 10, 11, 16], such as Theorem 3.1 and Theorem 3.2 in [6], Theorem 1 in [11]. The key assumption in [6] that the almost orbit u() is almost asymptotically isometric is not necessary in our theorems.

Theorem 4.3 Let X be a uniformly convex Banach space, and let C be a nonempty bounded closed convex subset of X. Let ={T(t):tG} of a reversible semigroup of asymptotically nonexpansive mappings on C, and let u() be an almost orbit of . If

ω- lim t G u(ht)u(t)=0

for every hG, then

ω ω (u)F().

Proof For any given ε(0,1], there exists t 0 G such that for any t t 0 , φ(t)< a ( ε ) 4 . Let p ω ω (u), then there exists a subnet { u ( t α ) } α A in { u ( t ) } t G with ω- lim α A u( t α )=p such that for all αA, t α t 0 , where A is a directed system. By Lemma 3.5, for any h G a ( a ( a ( ε ) 16 ) ) , there exists a nN such that for all αA,

1 n i = 1 n T ( h i ) u( t α ) F a ( ε ) 4 ( T ( h ) ) .

Noting for each αA,

1 n i = 1 n T ( h i ) u ( t α ) 1 n i = 1 n u ( h i t α ) φ( t α )< a ( ε ) 4 ,

we get

1 n i = 1 n u ( h i t α ) F a ( ε ) 4 ( T ( h ) ) +B ( 0 , a ( ε ) 4 ) F a ( ε ) ( T ( h ) ) .

Since u(ht)u(t)0 for every hG, we have u( h i t α )p, i=1,,n. Then for all h G a ( a ( a ( ε ) 16 ) ) ,

p=ω- lim α A 1 n i = 1 n u ( h i t α ) co ¯ F a ( ε ) ( T ( h ) ) .

Consequently, from Lemma 3.3, we can conclude that for all h G a ( a ( a ( ε ) 16 ) ) , p F ε (T(h)), which implies pF(). This completes the proof. □

Remark 4.3 In Theorem 4.1, Theorem 4.2 and Theorem 4.3, we do not assume that X has a Fréchet differentiable norm.

Theorem 4.4 Let X be a uniformly convex Banach space whose dual has the Kadec-Klee property, and let C be a nonempty bounded closed convex subset of X. Let ={T(t):tG} of a reversible semigroup of asymptotically nonexpansive mappings on C and u() be an almost orbit of . Then the following statements are equivalent:

  1. (1)

    ω ω (u)F().

  2. (2)

    ω- lim t G u(t)=pF()

  3. (3)

    ω- lim t G u(ht)u(t)=0 for every hG.

Proof (1) (2). It suffices to show that ω ω (u) is a singleton. Since X is reflexive, it is nonempty. Let f and g be two elements in ω ω (u), then by (1), we can obtain f,gF(). For any λ(0,1), by Lemma 3.6, lim t G λu(t)+(1λ)fg exists. Setting

h(λ)= lim t G λ u ( t ) + ( 1 λ ) f g ,

then for a given ε>0, there exists t 1 G such that for all t t 1 ,

λ u ( t ) + ( 1 λ ) f g h(λ)+ε.

Hence for all t t 1 ,

λ u ( t ) + ( 1 λ ) f g , j ( f g ) fg ( h ( λ ) + ε ) ,

where j(fg)J(fg). Let us note f co ¯ {u(t):t t 1 }, then

λ f + ( 1 λ ) f g , j ( f g ) fg ( h ( λ ) + ε ) ,

which means fgh(λ)+ε. Since ε is arbitrary, we get

fgh(λ).

It then follows from g ω ω (u) that there exists a subnet { u ( t α ) } α A in { u ( t ) } t G such that ω- lim α A u( t α )=g, where A is a directed system. Putting

I=A×N= { β = ( α , n ) : α A , n N } ,

then for β i =( α i , n i )I, i=1,2, we define β 1 β 2 if and only if α 1 α 2 , n 1 n 2 . In this case, I is also a directed system. For arbitrary β=(α,n)I, define P 1 β=α, P 2 β=n, t β = t α , ε β = 1 P 2 β , then ω- lim β I u( t β )=g and lim β I ε β =0. By Lemma 3.1, we have

λ u ( t β ) + ( 1 λ ) f g f g 2 +2λ u ( t β ) f , j ( λ u ( t β ) + ( 1 λ ) f g ) .

Applying Lemma 3.6 and noting the inequality fgh(λ), we obtain

lim inf β I u ( t β ) f , j ( λ u ( t β ) + ( 1 λ ) f g ) 0.

Then for each γI, there exists β γ I such that β γ γ and

u ( t β γ ) f , j ( ε γ u ( t β γ ) + ( 1 ε γ ) f g ) ε γ .
(4.2)

Obviously, { β γ } is also a subnet of I, then ω- lim γ I u( t β γ )=g. Putting

j γ =j ( ε γ u ( t β γ ) + ( 1 ε γ ) f g ) .

In as much as X is reflexive, X is also reflexive, we can conclude that the set of all weak limit points of { j γ ,γI} is nonempty. Hence, without loss of generality, we may assume that ω- lim γ I j γ =j X . Therefore, j lim inf γ I j γ =fg. Since

fg, j γ = ε γ u ( t β γ ) + ( 1 ε γ ) f g 2 ε γ u ( t β γ ) f , j γ ,

passing the limit for γI, we get fg,j= f g 2 , which implies jfg. Hence we can get

fg,j= f g 2 = j 2 ,

i.e., jJ(fg). Therefore, ω- lim γ I j γ =j and lim γ I j γ =j. Since X is reflexive and has Kadec-Klee property, it has the Kadec property, and this implies that lim γ I j γ =j. Taking the limit for γI in (4.2), we obtain gf,j0, i.e., f g 2 0, which implies f=g.

(2) (3). Obviously.

(3) (1). See Theorem 4.3. This completes the proof. □

Remark 4.4 By Theorem 4.4, we can get many known theorems in [2, 6, 10, 11, 1416], such as Theorem 4.3 and Theorem 8.1 in [14], Theorem 3.1 and Theorem 3.2 in [15]. And in [6, 10], it is assumed that the almost orbit u() is almost asymptotically isometric and the subspace D has a left invariant mean. Those key conditions are not necessary by the theorem above.

References

  1. Baillon JB: Un théorème de type ergodique pour les contraction non linéaires dans un espace de Hilbert. C. R. Math. Acad. Sci. Paris 1976, 280: 1511–1514.

    MathSciNet  Google Scholar 

  2. Takahashi W: A nonlinear ergodic theorem for a reversible semigroup of nonexpansive mappings in a Hilbert space. Proc. Am. Math. Soc. 1986, 17: 55–58.

    Article  Google Scholar 

  3. Li G, Kim JK: Nonlinear ergodic theorems for general curves defined on general semigroups. Nonlinear Anal. 2003, 55: 1–14. 10.1016/S0362-546X(03)00194-9

    Article  MathSciNet  Google Scholar 

  4. Li G, Kim JK: Asymptotic behavior of non-autonomous dissipative systems in Hilbert spaces. Acta Math. Sci. 1998, 18: 25–30.

    Google Scholar 

  5. Lau AT, Shioji N, Takahashi W: Existences of nonexpansive retractions for amenable semigroups of nonexpansive mappings and nonlinear ergodic theorems in Banach spaces. J. Funct. Anal. 1999, 161: 62–75. 10.1006/jfan.1998.3352

    Article  MathSciNet  Google Scholar 

  6. Kim JK, Li G: Nonlinear ergodic theorems for reversible semigroups of asymptotically nonexpansive mappings in Banach spaces. Dyn. Syst. Appl. 2000, 9: 255–268.

    MathSciNet  Google Scholar 

  7. Kaczor W, Kuczumow T, Michalska M: Convergence of ergodic means of orbits of semigroups of nonexpansive mappings in sets with the Γ-Opial property. Nonlinear Anal. 2007, 67: 2122–2130. 10.1016/j.na.2006.08.040

    Article  MathSciNet  Google Scholar 

  8. Kim KS: Nonlinear ergodic theorems of nonexpansive type mappings. J. Math. Anal. Appl. 2009, 358: 261–272. 10.1016/j.jmaa.2009.04.045

    Article  MathSciNet  Google Scholar 

  9. Li G, Kim JK: Nonlinear ergodic theorem for commutative semigroups of non-Lipschitzian mappings in Banach space. Houst. J. Math. 2003, 29: 23–36.

    MathSciNet  Google Scholar 

  10. Li G: Weak convergence and nonlinear ergodic theorem for reversible semigroups of non-Lipschitzian mappings. J. Math. Anal. Appl. 1997, 206: 451–464. 10.1006/jmaa.1997.5237

    Article  MathSciNet  Google Scholar 

  11. Oka H: Nonlinear ergodic theorems for commutative semigroups of asymptotically nonexpansive mappings. Nonlinear Anal. 1992, 7: 619–635.

    Article  MathSciNet  Google Scholar 

  12. Saeidi S: Ergodic retractions for amenable semigroups in Banach spaces with normal structure. Nonlinear Anal. 2009, 71: 2558–2563. 10.1016/j.na.2009.01.091

    Article  MathSciNet  Google Scholar 

  13. Saeidi S: Existence of ergodic retractions for semigroups in Banach spaces. Nonlinear Anal. 2008, 69: 3417–3422. 10.1016/j.na.2007.09.031

    Article  MathSciNet  Google Scholar 

  14. Falset JG, Kaczor W, Kuczumow T, et al.: Weak convergence theorems for asymptotically nonexpansive mappings and semigroups. Nonlinear Anal. 2001, 43: 377–401. 10.1016/S0362-546X(99)00200-X

    Article  MathSciNet  Google Scholar 

  15. Kaczor W: Weak convergence of almost-orbit for asymptotically nonexpansive commutative semigroups. J. Math. Anal. Appl. 2002, 272: 565–574. 10.1016/S0022-247X(02)00175-0

    Article  MathSciNet  Google Scholar 

  16. Miyadera I, Kobayasi K: On the asymptotic behavior of almost-orbits of nonlinear contraction semigroups in Banach space. Nonlinear Anal. 1982, 6: 349–365. 10.1016/0362-546X(82)90021-9

    Article  MathSciNet  Google Scholar 

  17. Aksoy AG, Khamsi MA: Nonstandard Methods in Fixed Point Theory. Springer, New York; 1990.

    Book  Google Scholar 

  18. Day MM: Amenable semigroups. Ill. J. Math. 1957, 1: 509–544.

    Google Scholar 

  19. Holmes RD, Narayanaswamy PP: On asymptotically nonexpansive semigroups of mappings. Can. Math. Bull. 1970, 13: 209–214. 10.4153/CMB-1970-042-1

    Article  MathSciNet  Google Scholar 

  20. Holmes RD, Lau AT: Asymptotically non-expansive actions of topological semigroups and fixed points. Bull. Lond. Math. Soc. 1971, 3: 343–347. 10.1112/blms/3.3.343

    Article  MathSciNet  Google Scholar 

  21. Kirk W, Torrejon R: Asymptotically nonexpansive semigroup in Banach space. Nonlinear Anal. 1979, 3: 111–121. 10.1016/0362-546X(79)90041-5

    Article  MathSciNet  Google Scholar 

  22. Saeidi S: Strong convergence of Browder’s type iterations for left amenable semigroups of Lipschitzian mappings in Banach spaces. J. Fixed Point Theory Appl. 2009, 5: 93–103. 10.1007/s11784-008-0092-3

    Article  MathSciNet  Google Scholar 

  23. Saeidi S: Approximating common fixed points of Lipschitzian semigroup in smooth Banach spaces. Fixed Point Theory Appl. 2008., 2008: Article ID 363257

    Google Scholar 

  24. Chang SS: Some problems and results in the study of nonlinear analysis. Nonlinear Anal. 1997, 30: 4197–4208. 10.1016/S0362-546X(97)00388-X

    Article  MathSciNet  Google Scholar 

  25. Bruck RE: A simple proof of the mean ergodic theorem for nonlinear contractions in Banach spaces. Isr. J. Math. 1979, 32: 107–116. 10.1007/BF02764907

    Article  MathSciNet  Google Scholar 

Download references

Acknowledgements

This research is supported by the National Science Foundation of China (11201410, 11271316 and 11101353), the Natural Science Foundation of Jiangsu Province (BK2012260) and the Natural Science Foundation of Jiangsu Education Committee (10KJB110012 and 11KJB110018).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Qianglian Huang.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

All authors contributed equally and significantly in writing this paper. All authors read and approved the final manuscript.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 2.0 International License ( https://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and permissions

About this article

Cite this article

Zhu, L., Huang, Q. & Li, G. Nonlinear ergodic theorems and weak convergence theorems for reversible semigroup of asymptotically nonexpansive mappings in Banach spaces. Fixed Point Theory Appl 2013, 231 (2013). https://doi.org/10.1186/1687-1812-2013-231

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/1687-1812-2013-231

Keywords