Journal of Applied Mathematics & Bioinformatics, vol.3, no.4, 2013, 285-300 ISSN: 1792-6602 (print), 1792-6939 (online) Scienpress Ltd, 2013 A New Iterative Scheme for Approximating Common Fixed Points of Two Nonexpansive Mappings in Banach Space Jinzuo Chen1 and Dingping Wu2 Abstract In this paper, we introduce the modified iterations of Mann’s type for nonexpansive mappings to have the strong convergence in a uniformly convex Banach space. We study approximation of common fixed point of nonexpansive mappings in Banach space by using a new iterative scheme. Mathematics Subject Classification: 47H09; 47H10; 47J25 Keywords: Strong convergence; Modified Mann iteration; Uniformly convex Banach space; Nonexpansive mapping; Common fixed point 1 2 College of Applied Mathematics, Chengdu University of Information Technology. E-mail: chanjanegeoger@163.com. College of Applied Mathematics, Chengdu University of Information Technology. E-mail: wdp68@163.com Article Info: Received : July 22, 2013. Revised : September 11, 2013. Published online : December 10, 2013. 286 1 A New Iterative Scheme for Approximating Common Fixed Points... Introduction Let E be a real Banach space, C a nonempty closed convex subset of E, and T : C → C a mapping. Recall that T is nonexpansive mapping[1] if kT x − T yk ≤ kx − yk for all x, y ∈ C. A point x ∈ C is a f ixed point of T provided T x = x. Denote by F ix(T ) the set of fixed point of T ; that is, F ix(T ) = {x ∈ C : T x = x}. It is assumed throughout the paper that T is a nonexpansive mapping such that F ix(T ) 6= ∅. Iterative methods are often used to solve the fixed point equation T x = x. The most well-known method is perhaps the Picard successive iteration method when T is a contraction. Picard’s method generates a sequence {xn } successively as xn = T xn−1 for n ≥ 1 with x0 arbitrary, and this sequence converges in norm to the unique fixed point of T . However, if T is not a contraction (for instance, if T is nonexpansive), then Picard’s successive iteration fail, in general, to converge. Instead, Mann’s iteration method [10] prevails. Mann’s method, an averaged process in nature, generates a sequence {xn } recursively via xn+1 = αn xn + (1 − αn )T xn , n ≥ 0, (1.1) where the initial guess x0 ∈ C is arbitrarily chosen. Mann’s iteration method (1.1) has been proved to be a powerful method for solving nonlinear operator equations involving nonexpansive mapping, asymptotically nonexpansive mapping, and other kinds of nonlinear mapping; see [2, 3, 6-8, 10, 12, 13, 15, 16] and the references therein. It is known that Mann’s iteration method (1.1) is in general not strongly convergent [5] for nonexpansive mappings. So to get strong convergence, one has to modify the iteration method (1.1). In this regard, we will show in Section 3. Motivated and inspired by the research going on in these fields, we suggest and analyze now new modified Mann’s iteration for finding the common fixed point of the nonexpansive mappings in Banach space. We propose the modified Mann’s iteration and consider the strong convergence of the approximate solutions for nonexpansive in Banach space. 287 Jinzuo Chen and Dingping Wu We suggest and analyze the following two iterative methods: x0 ∈ C chosen arbitrarily, y = β x + (1 − β )T x , n n n n n zn = γn xn + (1 − γn )Sxn , xn+1 = αn yn + (1 − αn )zn , n ≥ 0, (1.2) and if there exists two sequences {x0n } and {x00n } generated by x0 ∈ C chosen arbitrarily, 0 0 yn = βn xn + (1 − βn )T xn , zn = γn x00n + (1 − γn )Sx00n , x0n+1 = αn zn + (1 − αn )yn , 00 xn+1 = αn yn + (1 − αn )zn , n ≥ 0, (1.3) where αn , βn , γn are constants satisfying certain conditions. We write xn → x to indicate that the sequence {xn } converges strongly to x. We use F to denote the set of common fixed point of the mappings T and S. 2 Preliminaries This section collects some lemmas which will be used in the proofs for the main results in the next section. Lemma 2.1. [12] Let {an }, {bn } and {δn } be sequences of nonnegative real numbers satisfying the inequality an+1 ≤ (1 + δn )an + bn , n ≥ 1. If ∞ P δn < ∞ and n=1 (1) lim an exists; ∞ P bn < ∞, then n=1 n→∞ (2) lim an = 0 whenever lim inf an = 0. n→∞ n→∞ 288 A New Iterative Scheme for Approximating Common Fixed Points... Lemma 2.2. [14] Suppose that E is a uniformly convex Banach space and 0 < tn < 1 for all n ∈ N . Let {xn } and {yn } be two sequences of E such that lim sup kxn k ≤ r, lim sup kyn k ≤ r and lim ktn xn + (1 − tn )yn k = r hold for n→∞ n→∞ n→∞ some r ≥ 0, then lim kxn − yn k = 0. n→∞ Lemma 2.3. [15] A mapping T :C→C with nonempty fixed point set F in C will be said to satisfy Condition (I): If there is a nondecreasing function f : [0, ∞) → [0, ∞) with f (0) = 0, f (r) > 0 for all r ∈ (0, ∞) such that kx − T xk ≥ f (d(x, F )) f or all x ∈ C, where d(x, F ) = inf{kx − pk : p ∈ F }. Lemma 2.4. [18] Given a number r > 0. A real Banach space E is uniformly convex if and only if there exists a continuous strictly increasing function ϕ : [0, ∞] → [0, ∞], ϕ(0) = 0, such that kλx + (1 − λ)yk2 ≤ λkxk2 + (1 − λ)kyk2 − λ(1 − λ)ϕ(kx − yk) for all λ ∈ [0, 1] and x, y ∈ E such that kxk ≤ r and kyk ≤ r. Note that the inequality in Lemma 2.4 is known as Xu0 s inequality. It follows from Lemma 2.4 that we get the following lemma can be found in [4]. Lemma 2.5. Given a number r > 0. Let E is a uniformly convex Banach space then there exists a continuous strictly increasing function ϕ : [0, ∞] → [0, ∞] with ϕ(0) = 0, such that kλx + µy + γzk2 ≤ λkxk2 + µkyk2 + γkzk2 − λµϕ(kx − yk) for all λ, µ, γ ∈ [0, 1] and x, y, z ∈ E such that kxk ≤ r, kyk ≤ r and kzk ≤ r. Lemma 2.6. [9] Let {αn }, {βn } be sequences of nonnegative real numbers ∞ ∞ P P such that αn = ∞. If αn βn < ∞, then lim inf βn = 0. n=1 n=1 n→∞ 289 Jinzuo Chen and Dingping Wu 3 Convergence to a common fixed point of nonexpansive mappings 3.1 There exists one sequence {xn } In this part, we prove our main theorem for finding a common fixed point of nonexpansive mappings in Banach space in the case of one sequence. Theorem 3.1. Let C be a nonempty closed convex subset of a uniformly convex Banach space E and let T and S be two nonexpansive self mappings of C satisfy Condition (I) and F 6= ∅. Given {αn }, {βn } and {γn } are sequences in (0,1), P P P such that αn < ∞, γn βn = ∞, (1 − γn ) < ∞ for all n ≥ 1. ∞ Define a sequence {xn }∞ n=0 in C by the algorithm (1.2), then {xn }n=0 strongly converges to a common fixed point of T and S. Proof. First, we observe that {xn } is bounded, if we take an arbitrary fixed point q of F , noting that kyn − qk ≤ kxn − qk and kzn − qk ≤ kxn − qk, we have kxn+1 − qk = kαn yn + (1 − αn )zn − qk ≤ αn kyn − zn k + kzn − qk ≤ αn kyn − qk + αn kzn − qk + kzn − qk By Lemma 2.1 and P (3.1) ≤ (1 + 2αn )kxn − qk. αn < ∞, thus lim kxn − qk exists. Denote n→∞ lim kxn − qk = c. n→∞ Hence, {xn } is bounded, so are {yn } and {zn }. Now kxn+1 − qk = kαn yn + (1 − αn )zn − qk = kαn (yn − zn ) + (zn − q)k ≤ αn kyn − zn k + kzn − qk. By P αn < ∞ and the boundedness of {zn } and {yn }, we obtain lim kxn − qk ≤ lim inf kzn − qk. n→∞ n→∞ (3.2) Since kzn − qk ≤ kxn − qk, which implies that lim sup kzn − qk ≤ lim kxn − qk, n→∞ n→∞ (3.3) 290 A New Iterative Scheme for Approximating Common Fixed Points... so that (3.2) and (3.3) give lim kzn − qk = lim kxn − qk = c. n→∞ n→∞ Moreover, kSxn − qk ≤ kxn − qk implies that lim sup kSxn − qk ≤ c. n→∞ Thus c = lim kzn − qk = lim kγn xn + (1 − γn )Sxn − qk n→∞ n→∞ = lim kγn (xn − q) + (1 − γn )(Sxn − q)k, n→∞ given by Lemma 2.2 that lim kSxn − xn k = 0. n→∞ By (3.1) and P (3.4) αn < ∞, then we have kxn+m − qk ≤ (1 + 2αn+m−1 )kxn+m−1 − qk ≤ e2αn+m−1 kxn+m−1 − qk ≤ e2αn+m−1 e2αn+m−2 kxn+m−2 − qk ≤ ... 2 n+m−1 P ≤e i=n αi kxn − qk. That is kxn+m − qk ≤ M kxn − qk, 2 n+m−1 P (3.5) αi where M = e i=n , for all m, n ≥ 1, for all q ∈ F and for M > 0. Next we prove that {xn }∞ n=0 is Cauchy sequence. Since q ∈ F arbitrarily and lim kxn − qk exists, consequently d(xn , F ) n→∞ exists by Lemma 2.3. From the Lemma 2.3 and (3.4), we get lim f (d(xn , F )) ≤ lim kxn − Sxn k = 0 n→∞ n→∞ Since f : [0, ∞) → [0, ∞) is a nondecreasing function satisfy f (0) = 0, f (r) > 0 for all r ∈ (0, ∞), therefore we have lim d(xn , F ) = 0. n→∞ 291 Jinzuo Chen and Dingping Wu Let ε > 0, since lim d(xn , F ) = 0 and therefore exists a constant n0 such n→∞ that for all n ≥ n0 , we have d(xn , F ) ≤ ε , 2M d(xn0 , F ) ≤ ε . 2M in particular, There must exist p1 ∈ F , such that d(xn0 , p1 ) ≤ ε . 2M From (3.5), it can be obtained that when n ≥ n0 , kxn+m − xn k ≤ kxn+m − p1 k + kxn − p1 k ≤ 2M kxn0 − p1 k ε = ε. ≤ 2M · 2M This implies {xn }∞ n=0 is a Cauchy sequence in a closed convex subset C of a Banach space E. Thus, it must converges to a point in C, let lim xn = p. n→∞ For all ² > 0, as lim xn = p, thus there exists a number n1 such that when n→∞ n2 ≥ n1 , ² kxn2 − pk ≤ . (3.6) 4 In fact, lim d(xn , F ) = 0 implies that using the n2 above, when n ≥ n2 , n→∞ we get ² d(xn , F ) ≤ . 8 In particular, d(xn2 , F ) ≤ 8² . Thus, there must exist p̄ ∈ F , such that ² kxn2 − p̄k = d(xn2 , p̄) = . 8 From (3.6) and (3.7), we get kSp − pk = kSp − p̄ + Sxn2 − p̄ + p̄ − xn2 + xn2 − p + p̄ − Sxn2 k ≤ kSp − p̄k + kxn2 − p̄k + kxn2 − pk + 2kSxn2 − p̄k ≤ kp − p̄k + 3kxn2 − p̄k + kxn2 − pk ≤ kxn2 − pk + kxn2 − p̄k + 3kxn2 − p̄k + kxn2 − pk = 4kxn2 − p̄k + 2kxn2 − pk 4² 2² + = ². ≤ 8 4 (3.7) 292 A New Iterative Scheme for Approximating Common Fixed Points... As ² is an arbitrary positive number, thus, Sp = p. Let un+1 = γn xn + βn T xn + (1 − βn − γn )T xn . Then we have kun+1 − qk2 = kγn xn + βn T xn + (1 − βn − γn )T xn − qk2 = kγn (xn − q) + βn (T xn − q) + (1 − βn − γn )(T xn − q)k2 ≤ γn kxn − qk2 + βn kT xn − qk2 + (1 − βn − γn )kT xn − qk2 − γn βn ϕkxn − T xn k ≤ kxn − qk2 − γn βn ϕkxn − T xn k, and hence γn βn ϕkxn − T xn k ≤ kxn − qk2 − kun+1 − qk2 for q ∈ F . Summing from n=1 to ∞, we have ∞ X γn βn ϕkxn − T xn k = ∞ X (kxn − qk2 − kun+1 − qk2 ) n=1 n=1 = ∞ X (kxn − qk + kun+1 − qk)(kxn − qk − kun+1 − qk) n=1 ≤ 2K ≤ 2K ≤ 4K ∞ X kun+1 − xn k n=1 ∞ X (1 − γn )kxn − T xn k n=1 ∞ X 2 (1 − γn ), n=1 where K = sup{kxn − qk}, from n∈ N ∞ X ∞ P n=1 (1 − γn ) < ∞, we get n=1 ∞ X γn βn ϕkxn − T xn k ≤ 4K (1 − γn ) < ∞. 2 n=1 Since ∞ P n=1 γn βn = ∞, from Lemma 2.6, we get lim inf ϕkT xn − xn k = 0. Hence n→∞ lim inf kT xn − xn k = 0. n→∞ (3.8) 293 Jinzuo Chen and Dingping Wu Since T is a nonexpansive mapping, we have kT xn+1 − xn+1 k = kT xn+1 − αn yn − (1 − αn )zn k = kT xn+1 − T xn + T xn + αn T xn − αn T xn − αn yn − (1 − αn )zn k ≤ kT xn+1 − T xn k + (1 − αn )kT xn − zn k + αn kT xn − yn k ≤ kxn+1 − xn k + (1 − αn )kT xn − zn k + αn kT xn − yn k = kαn yn + (1 − αn )zn − xn k + (1 − αn )kT xn − zn k + αn kT xn − yn k ≤ αn kyn − xn k + (1 − αn )kzn − xn k + (1 − αn )kT xn − zn k + αn kT xn − yn k = αn kβn xn + (1 − βn )T xn − xn k + (1 − αn )kγn xn + (1 − γn )Sxn − xn k + (1 − αn )kT xn − γn xn − (1 − γn )Sxn k + αn kT xn − βn xn − (1 − βn )T xn k ≤ αn (1 − βn )kT xn − xn k + (1 − αn )(1 − γn )kSxn − xn k + αn βn kT xn − xn k + γn (1 − αn )kT xn − xn k + (1 − αn )(1 − γn )kT xn − Sxn k ≤ αn (1 − βn )kT xn − xn k + (1 − γn )kSxn − xn k + αn βn kT xn − xn k + (1 − αn )kT xn − xn k + (1 − γn )kT xn − Sxn k ≤ kT xn − xn k + (1 − γn )(kSxn − xn k + kT xn − Sxn k). Since ∞ P (1 − γn ) < ∞, it follows from Lemma 2.1 that lim kT xn − xn k exists. n→∞ n=1 Therefore, from (3.8), we get lim kT xn − xn k = 0. n→∞ Then using the same argument we can show that {xn }∞ n=0 converges strongly to a common fixed point of T and S. Remark 3.1. It is to be noted that (1.2) reduces to (1.1) when T or S equal to I. 3.2 There exists two sequences {x0n } and {x00n } In this part, we prove our main theorem for finding a common fixed point of nonexpansive mappings in Banach space in the case of two sequences. 294 A New Iterative Scheme for Approximating Common Fixed Points... Theorem 3.2. Let C be a nonempty bounded closed convex subset of a uniformly convex Banach space E and let T and S be two nonexpansive self mappings of C satisfy Condition (I) and F 6= ∅. Given {αn }, {βn } and {γn } are P P sequences in (0,1), such that αn < ∞, βn < ∞, βn > γn for all n ≥ 1. 0 ∞ 00 ∞ Define two sequences {xn }n=0 and {xn }n=0 in C by the algorithm (1.3), 00 ∞ then {x0n }∞ n=0 and {xn }n=0 strongly converge to a common fixed point of T and S. Proof. By the boundedness of C, we obverse that both {x0n } and {x00n } are bounded, if we take an arbitrary fixed point q of F , noting that kyn − qk ≤ kx0n − qk and kzn − qk ≤ kx00n − qk, we have kx0n+1 − qk = kαn zn + (1 − αn )yn − qk ≤ αn kzn − yn k + kyn − qk ≤ αn kyn − qk + αn kzn − qk + kyn − qk (3.9) ≤ (1 + αn )kx0n − qk + αn kx00n − qk. By Lemma 2.1 and P αn < ∞, thus lim kx0n − qk exists. Denote n→∞ lim kx0n − qk = c0 . n→∞ Similarly, we have lim kx00n − qk = c00 . n→∞ Since both {x0n } and {x00n } are bounded, we get {yn } and {zn } are bounded. Now kx0n+1 − qk = kαn zn + (1 − αn )yn − qk = kαn (zn − yn ) + (yn − q)k ≤ αn kzn − yn k + kyn − qk. By P αn < ∞, we obtain lim kx0n − qk ≤ lim inf kyn − qk. n→∞ n→∞ (3.10) Since kyn − qk ≤ kx0n − qk, which implies that lim sup kyn − qk ≤ lim kx0n − qk, n→∞ n→∞ (3.11) 295 Jinzuo Chen and Dingping Wu so that (3.10) and (3.11) give lim kyn − qk = lim kx0n − qk = c0 . n→∞ n→∞ Moreover, kT x0n − qk ≤ kx0n − qk implies that lim sup kT x0n − qk ≤ c0 . n→∞ Thus c0 = lim kyn − qk = lim kβn x0n + (1 − βn )T x0n − qk n→∞ n→∞ = lim kβn (x0n − q) + (1 − βn )(T x0n − q)k, n→∞ given by Lemma 2.2 that lim kT x0n − x0n k = 0. (3.12) n→∞ By (3.9) and P αn < ∞, then we have kx0n+m − qk ≤ (1 + αn+m−1 )kx0n+m−1 − qk + αn+m−1 kx00n+m−1 − qk ≤ eαn+m−1 kx0n+m−1 − qk + αn+m−1 kx00n+m−1 − qk ≤ eαn+m−1 eαn+m−2 kxn+m−2 − qk + eαn+m−1 (αn+m−2 kx00n+m−2 − qk + αn+m−1 kx00n+m−1 − qk) ≤ ... n+m−1 P ≤e i=n αi n+m−1 P kx0n − qk + e i=n αi · n+m−1 X αi kx00i − qk. i=n That is kx0n+m − qk ≤ L(kx0n − qk + n+m−1 X si ), (3.13) i=n n+m−1 P αi where L = e i=n , si = αi kx00i − qk. for all m, n ≥ 1, for all q ∈ F . Next we prove that {x0n }∞ n=0 is Cauchy sequence. Since q ∈ F arbitrarily and lim kx0n − qk exists, consequently d(x0n , F ) n→∞ exists by Lemma 2.3. From the Lemma 2.3 and (3.12), we get lim f (d(x0n , F )) ≤ lim kx0n − T x0n k = 0 n→∞ n→∞ 296 A New Iterative Scheme for Approximating Common Fixed Points... Since f : [0, ∞) → [0, ∞) is a nondecreasing function satisfy f (0) = 0, f (r) > 0 for all r ∈ (0, ∞), therefore we have lim d(x0n , F ) = 0. n→∞ Let ε > 0, since lim d(x0n , F ) = 0 and n→∞ ∞ P si < ∞, therefore it exists a i=0 constant n0 such that for all n ≥ n0 , we have d(x0n , F ) ≤ ε , 2L d(x0n0 , F ) ≤ ε . 2L in particular, There must exist p1 ∈ F , such that d(x0n0 , p1 ) ≤ ε . 2L From (3.13), it can be obtained that when n ≥ n0 , kx0n+m − x0n k ≤ kx0n+m − p1 k + kx0n − p1 k ≤ 2Lkx0n0 − p1 k ε = ε. ≤ 2L · 2L This implies {x0n }∞ n=0 is a Cauchy sequence in a closed convex subset C of a Banach space E. Thus, it must converges to a point in C, let lim x0n = x0 . n→∞ For all ² > 0, as lim x0n = x0 , thus there exists a number n1 such that when n→∞ n2 ≥ n1 , ² (3.14) kx0n2 − x0 k ≤ . 4 In fact, lim d(x0n , F ) = 0 implies that using the n2 above, when n ≥ n2 , n→∞ we get ² d(x0n , F ) ≤ . 8 In particular, d(x0n2 , F ) ≤ 8² . Thus, there must exist x̄0 ∈ F , such that ² kx0n2 − x̄0 k = d(x0n2 , x̄0 ) = . 8 (3.15) 297 Jinzuo Chen and Dingping Wu From (3.14) and (3.15), we get kT x0 − x0 k = kT x0 − x̄0 + T x0n2 − x̄0 + x̄0 − x0n2 + x0n2 − x0 + x̄0 − T x0n2 k ≤ kT x0 − x̄0 k + kx0n2 − x̄0 k + kx0n2 − x0 k + 2kT x0n2 − x̄0 k ≤ kx0 − x̄0 k + 3kx0n2 − x̄0 k + kx0n2 − x0 k ≤ kx0n2 − x0 k + kx0n2 − x̄0 k + 3kx0n2 − x̄0 k + kx0n2 − x0 k = 4kx0n2 − x̄0 k + 2kx0n2 − x0 k 4² 2² ≤ + = ². 8 4 As ² is an arbitrary positive number, thus, T x0 = x0 . Similarly, we have lim kSx00n −x00n k = 0, and then Sx00 = x00 (x00n → x00 as n → n→∞ ∞) Finally, we prove x0 = x00 . Let wn+1 = αn T x0n + (1 − αn )Sx00n and kx00n − qk ≥ kx0n − qk for all n ≥ 1. Then kwn+1 − qk = αn kx0n − qk + (1 − αn )kx00n − qk ≤ max{kx0n − qk, kx00n − qk} ≤ kx00n − qk. Now, kx00n+1 − qk = kαn yn + (1 − αn )zn − qk = kαn βn x0n + αn (1 − βn )T x0n + (1 − αn )γn x00n + (1 − αn )(1 − γn )Sx00n − qk ≤ βn kαn x0n + (1 − αn )x00n k + kαn T x0n + (1 − αn )Sx00n − qk = βn kαn x0n + (1 − αn )x00n k + kwn+1 − qk, since P βn < ∞ and the boundedness of {x0n } and {x00n }, we get c00 ≤ lim kwn+1 − qk. n→∞ Then we have lim kwn+1 − qk = c00 . n→∞ Moreover, kT x0n − qk ≤ kx0n − qk and kSx00n − qk ≤ kx00n − qk, imply that lim sup kT x0n − qk ≤ c00 and lim sup kSx00n − qk ≤ c00 . n→∞ n→∞ 298 A New Iterative Scheme for Approximating Common Fixed Points... Thus, c00 = lim kwn+1 − qk = kαn T x0n + (1 − αn )Sx00n − qk n→∞ = kαn (T x0n − q) + (1 − αn )(Sx00n − q)k, given by Lemma 2.2 that lim kSx00n − T x0n k = 0. n→∞ (3.16) So lim kx0n+1 − x00n+1 k n→∞ = lim (2αn − 1)kzn − yn k n→∞ = lim (2αn − 1)kγn (x00n − Sx00n ) − βn (x0n − T x0n ) + (Sx00n − T x0n )k, n→∞ so we obtain lim kx0n+1 − x00n+1 k = 0 for (3.12) and (3.16), it means x0 = x00 . n→∞ This completes the proof. Acknowledgements This work was supported by the National Natural Science Foundation of China(Grant No. 11171046) and supported by the Scientific Research Foundation of CUIT(No. KYTZ201004). References [1] F.E. Browder, Fixed point theorems for noncompact mappings in Hilbert space, Proc. Nat. Acad., 53, (1965), 1272-1276. [2] F.E. Browder and W.V. Petryshyn, Construction of fixed point of nonlinear mappings in Hilbert spaces, J. Math. Appl., 20, (1967), 197-228. [3] C. Byrne, A unified treatment of some iterative algorithms in signal processing and image constrution, Inverse Probl., 20, (2004), 103-120. Jinzuo Chen and Dingping Wu 299 [4] Y.J. Cho, H.Y. Zhou and G.T. Guo, Weak and strong convergence theorems for three-step iterations with errors for asymptotically nonexpansive mappings, Comput. Math. Appl., 47, (2004), 707-717. [5] A. Genel and J. Lindenstrauss, An example concerning fixed points, Israel J. Math., 22, (1975), 81-86. [6] B. Halpern, Fixed points of nonexpanding maps, Bull. Amer. Math. Soc., 73, (1967), 957-961. [7] T.H. Kim and H.K. Xu, Strong convergence of modified Mann iterations, Nonlinear Anal., 61, (2005), 51-60. [8] T.H. Kim and H.K. Xu, Strong convergence of modified Mann iterations for asymptotically nonexpansive mappings and semigroups, Nonlinear Anal., 64, (2006), 1140-1152. [9] K.S. Kim, Approximating common fixed points of nonspreading-type mappings and nonexpansive mappings in a Hilbert space, Abstract and Applied Analysis, 2012(2012), doi: 10.1155/2012/594218 [10] W.R. Mann, Mean value methods on iteration, Proc. Amer. Math. Soc., 4, (1953), 506-510. [11] K. Nakajo and W. Takahashi, Strong convergence theorems for nonexpansive mappings and nonexpansive semigroups, J. Math. Anal. Appl., 279, (2003), 372-379. [12] K. Nammanee, M.A. Noor and S. Suantai, Convergence criteria of modified Noor iterations with errors for asymptotically nonexpansive mappings, J. Math. Anal. Appl., 314, (2006), 320-334. [13] S. Reich, Weak convergence theorems for nonexpansive mappings in Banach spaces, J. Math. Anal. Appl., 67, (1979), 274-276. [14] J. Schu, Weak and strong convergence to fixed points of asymptotically nonexpansive mappings, Bull. Austral. Math. Soc., 43, (1991), 153-159. [15] H.F. Senter and W.G. Dotson, Jr, Approximating fixed points of nonexpansive mappings, Proc. Amer. Math. Soc., 44, (1974), 375-380. 300 A New Iterative Scheme for Approximating Common Fixed Points... [16] H.K. Xu, Iterative algorithms for nonlinear operators, J. Lond. Math. Soc., 66, (2002), 240-256. [17] H.K. Xu, Strong convergence of an iterative method for nonexpansive and accretive operators, J. Math. Anal. Appl., 314, (2006), 631-643. [18] H.K. Xu, Inequalities in Banach spaces with applications, Nonlinear Anal., 16, (1991), 1127-1138.