() @ Appl. Gen. Topol. 19, no. 1 (2018), 65-84doi:10.4995/agt.2018.7677 c© AGT, UPV, 2018 Relation-theoretic metrical coincidence and common fixed point theorems under nonlinear contractions Md Ahmadullah, Mohammad Imdad and Mohammad Arif Department of Mathematics, Aligarh Muslim University, Aligarh,-202002, U.P., India (ahmadullah2201@gmail.com, mhimdad@gmail.com, mohdarif154c@gmail.com) Communicated by S. Romaguera Abstract In this paper, we prove coincidence and common fixed points results under nonlinear contractions on a metric space equipped with an arbi- trary binary relation. Our results extend, generalize, modify and unify several known results especially those are contained in Berzig [J. Fixed Point Theory Appl. 12, 221-238 (2012))] and Alam and Imdad [To ap- pear in Filomat (arXiv:1603.09159 (2016))]. Interestingly, a corollary to one of our main results under symmetric closure of a binary relation remains a sharpened version of a theorem due to Berzig. Finally, we use examples to highlight the accomplished improvements in the results of this paper. 2010 MSC: 47H10; 54H25. Keywords: complete metric spaces; binary relations; contraction mappings; fixed point. 1. Introduction Banach contraction principle (see [8]) continues to be one of the most in- spiring and core result of metric fixed point theory which also has various applications in classical functional analysis besides several other domains espe- cially in mathematical economics and psychology. In the course of last several Received 14 May 2017 – Accepted 18 September 2017 http://dx.doi.org/10.4995/agt.2018.7677 M. Ahmadullah, M. Imdad and M. Arif years, numerous authors have extended this result by weakening the contrac- tion conditions besides enlarging the class of underlying metric space. In recent years such type of results are also established employing order-theoretic notions. Historically speaking, the idea of order-theoretic fixed points was initiated by Turinici [23] in 1986. In 2004, Ran and Reurings [21] formulated a relatively more natural order-theoretic version of classical Banach contraction principle. The existing literature contains several relation-theoretic results on fixed, co- incidence and common fixed point (e.g., partial order: Ran and Reurings [21] and Nieto and Rodŕıguez-López [20], tolerance: Turinici [25, 26], strict order: Ghods et al. [12], transitive: Ben-El-Mechaiekh [10], preorder: Turinici [24] etc). Berzig [9] established the common fixed point theorem for nonlinear con- traction under symmetric closure of a arbitrary relation. Most recently, Alam and Imdad [5] proved a relation-theoretic version of Banach contraction prin- ciple employing amorphous relation which in turn unify the several well known relevant order-theoretic fixed point theorems. Moreover, for further details one can consults [1, 2, 4, 5, 6, 10, 9, 11, 14, 21, 20, 22, 25, 26, 13]. Our aim in this work is to proved some coincidence and common fixed point theorems for nonlinear contraction on metric space endowed with amorphous relation. The results proved herein generalize and unify main results of Berzig [9], Alam and Imdad [5] and several others. To demonstrate the validity of the hypotheses and degree of generality of our results, we also furnish some examples. 2. Preliminaries For the sake of simplicity to have possibly self-contained presentation, we require some basic definitions, lemmas and propositions for our subsequent discussion. Definition 2.1 ([15, 16]). Let (f, g) be a pair of self-mappings defined on a non-empty set X. Then (i) a point u ∈ X is said to be a coincidence point of the pair (f, g) if fu = gu, (ii) a point v ∈ X is said to be a point of coincidence of the pair (f, g) if there exists u ∈ X such that v = fu = gu, (iii) a coincidence point u ∈ X of the pair (f, g) is said to be a common fixed point if u = fu = gu, (iv) a pair (f, g) is called commuting if f(gu) = g(fu), ∀ u ∈ X. (v) a pair (f, g) is said to be weakly compatible if f and g commutes at their coincidence points i.e., f(gu) = g(fu) whenever f(u) = g(u) for any u ∈ X. Definition 2.2 ([17, 28, 27]). Let (f, g) be a pair of self-mappings defined on a metric space (X, d). Then (i) (f, g) is said to be weakly commuting if for all u ∈ X, d(f(gu), g(fu)) ≤ d(fu, gu), c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 66 Relation-theoretic metrical coincidence and common fixed point theorems (ii) (f, g) is said to be compatible if limn→∞ d(f(gun), g(fun)) = 0 whenever {un} ⊂ X is a sequence such that limn→∞ gun = limn→∞ fun, (iii) f is said to be a g-continuous at u ∈ X if gun d−→ gu, for all sequence {un} ⊂ X, we have fun d−→ fu. Moreover, f is said to be a g-continuous if it is continuous at every point of X. Definition 2.3 ([18]). A subset R of X × X is called a binary relation on X. We say that “u relates v under R” if and only if (u, v) ∈ R. Throughout this paper, R stands for a ‘non-empty binary relation’ (i.e., R 6= ∅) instead of ‘binary relation’ and Rs := R ∪ R−1, while N0, Q and Qc stand the set of whole numbers (N0 = N ∪ {0}), the set of rational numbers and the set of irrational numbers respectively. Definition 2.4 ([19]). A binary relation R defined on a non-empty set X is called complete if every pair of elements of X are comparable under that relation i.e., for all u, v in X, either (u, v) ∈ R or (v, u) ∈ R which is denoted by [u, v] ∈ R. Proposition 2.5 ([4]). Let R be a binary relation defined on a non-empty set X. Then (u, v) ∈ Rs if and only if [u, v] ∈ R. Definition 2.6 ([4]). Let f be a self-mapping defined on a non-empty set X. Then a binary relation R on X is called f-closed if for all u, v ∈ X (u, v) ∈ R ⇒ (fu, fv) ∈ R. Definition 2.7 ([5]). Let (f, g) be a pair of self-mappings defined on a non- empty set X. Then a binary relation R on X is called (f, g)-closed if for all u, v ∈ X, (gu, gv) ∈ R ⇒ (fu, fv) ∈ R. Notice that on setting g = I, (the identity mapping on X) Definition 2.7 reduces to Definition 2.6. Definition 2.8 ([4]). Let R be a binary relation defined on a non-empty set X. Then a sequence {un} ⊂ X is said to be an R-preserving if (un, un+1) ∈ R, ∀ n ∈ N0. Definition 2.9 ([5]). Let (X, d) be a metric space equipped with a binary relation R. Then (X, d) is said to be an R-complete if every R-preserving Cauchy sequence in X converges to a point in X. Remark 2.10 ([5]). Every complete metric space is R-complete, where R de- notes a binary relation. Moreover, if R is universal relation, then notions of completeness and R-completeness are same. Definition 2.11 ([5]). Let (X, d) be a metric space equipped with a binary relation R. Then a mappings f : X → X is said to be an R-continuous at u if un d−→ u, for any R-preserving sequence {un} ⊂ X, we have fun d−→ fu. Moreover, f is said to be an R-continuous if it is R-continuous at every point of X. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 67 M. Ahmadullah, M. Imdad and M. Arif Definition 2.12 ([5]). Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R. Then f is said to be a (g, R)- continuous at u if gun d−→ gu, for any R-preserving sequence {un} ⊂ X, we have fun d−→ fu. Moreover, f is called a (g, R)-continuous if it is (g, R)- continuous at every point of X. Notice that on setting g = I (the identity mapping on X), Definition 2.12 reduces to Definition 2.11. Remark 2.13. Every continuous mapping is R-continuous, where R denotes a binary relation. Moreover, if R is universal relation, then notions of R- continuity and continuity are same. Definition 2.14. Let (X, d) be a metric space equipped with a binary relation R and g a self-mapping on X. Then (X, d) is said to be (g, R)-regular if for any R-preserving sequence {un} with un → u such that [gun, gu] ∈ R ∀ n ∈ N. Definition 2.15 ([4]). Let (X, d) be a metric space. Then a binary rela- tion R on X is said to be d-self-closed if for any R-preserving sequence {un} with un d−→ u, there is a subsequence {unk} of {un} such that [unk, u] ∈ R, for all k ∈ N. Definition 2.16 ([5]). Let g be a self-mapping on a metric space (X, d). Then a binary relation R on X is said to be (g, d)-self-closed if for any R-preserving sequence {un} with un d−→ u, there is a subsequence {unk} of {un} such that [gunk, gu] ∈ R, for all k ∈ N. Notice that under the consideration g = I (the identity mapping on X), Definition 2.16 turn out to be Definition 2.15. Definition 2.17 ([22]). Let (X, d) be a metric space endowed with an arbitrary binary relation R. Then a subset D of X is said to be an R-directed if for every pair of points u, v in D, there is w in X such that (u, w) ∈ R and (v, w) ∈ R. Definition 2.18 ([5]). Let g be a self-mapping on a metric space (X, d) en- dowed with a binary relation R. Then a subset D of X is said to be a (g, R)-directed if for every pair of points u, v in D, there is w in X such that (u, gw) ∈ R and (v, gw) ∈ R. Notice that on setting g = I (the identity mapping on X), Definition 2.18 turn out to be Definition 2.17. Definition 2.19 ([5]). Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R. Then the pair (f, g) is said to be an R-compatible if lim n→∞ d(g(fun), f(gun)) = 0, whenever lim n→∞ g(un) = lim n→∞ f(un), for any sequence {un} ⊂ X such that {fun} and {gun} are R- preserving. For a given non-empty set X, together with a binary relation R on X and a pair of self-mappings (f, g) on X, we use the following notations: c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 68 Relation-theoretic metrical coincidence and common fixed point theorems • C(f, g): the collection of all coincidence points of (f, g); • Mf (gu, gv) := max { d(gu, gv), d(gu, fu), d(gv, fv), 1 2 [d(gu, fv) + d(gv, fu)] } ; and • Nf (gu, gv) := max { d(gu, gv), 1 2 [d(gu, fu) + d(gv, fv)], 1 2 [d(gu, fv) + d(gv, fu)] } . Remark 2.20. Observe that, Nf (gu, gv) ≤ Mf (gu, gv), for all u, v ∈ X. Let Φ be the family of all mappings ϕ : [0, ∞) → [0, ∞) satisfying the following properties: (Φ1): ϕ is increasing; (Φ2): ∞ ∑ n=1 ϕn(t) < ∞ for each t > 0, where ϕn is the n-th iterate of ϕ. Lemma 2.21 ([22]). Let ϕ ∈ Φ. Then for all s > 0, we have ϕ(s) < s. Proposition 2.22. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R and ϕ ∈ Φ. Then the following conditions are equivalent: (I): d(fu, fv) ≤ ϕ(Mf (gu, gv)) with (gu, gv) ∈ R; (II): d(fu, fv) ≤ ϕ(Mf (gu, gv)) with [gu, gv] ∈ R. Proof. The implication (II) ⇒ (I) is straightforward. To show that (I) ⇒ (II), choose u, v ∈ X such that [gu, gv] ∈ R. If (gu, gv) ∈ R, then (II) immediately follows from (I). Otherwise, if (gv, gu) ∈ R, then by (I) and the symmetry of metric d, we obtained the conclusion. � For the sake of completeness, we state the following theorems: Theorem 2.23 ([9, Theorems 3.2]). Let (f, g) be a pair of self-mappings de- fined on a metric space (X, d) equipped with a symmetric closure S := R∪R−1 of any binary relation R. Suppose the following conditions hold: (a) (X, d) is complete; (b) there exists w0 ∈ X such that (gw0, fw0) ∈ S; (c) S is (f, g) closed; (d) (X, d, S) is regular; (e) there exists ϕ ∈ Φ such that d(fu, fv) ≤ ϕ(Nf (gu, gv)) for all u, v ∈ X with (gu, gv) ∈ S. Then (f, g) has a unique coincidence point. Moreover, if C(f, g) is (g, S)- directed and (f, g) is weakly compatible, then (f, g) has a unique common fixed point. Theorem 2.24 ([5, Theorem 2]). Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R and Y a subspace of X. Assume that the following conditions hold: (f) (Y, d) is R-complete subspace of X; (g) f(X) ⊆ Y ∩ g(X); (h) ∃ w0 ∈ X such that (gw0, fw0) ∈ R; c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 69 M. Ahmadullah, M. Imdad and M. Arif (i) R is (f, g)-closed; (j) there exists α ∈ [0, 1) such that d(fu, fv) ≤ αd(gu, gv) for all u, v ∈ X with (gu, gv) ∈ R; (k) (k1) Y ⊆ g(X); (k2) either f is (g, R)-continuous or f and g are continuous or R|Y is d-self-closed; or, alternatively (l) (l1) (f, g) is R-compatible; (l2) g is R-continuous; (l3) f is R-continuous or R is (g, d)-self-closed. Then (f, g) has a coincidence point. Indeed, the main results of this paper are based on the following points: • Theorem 2.23 is improved by replacing symmetric closure S of any binary relation with arbitrary binary relation R, • Theorems 2.23 (upto coincidence point) and 2.24 are unified by replac- ing more general contraction condition, • Theorem 2.23 is generalized by replacing comparatively weaker notions namely R-completeness of any subspace Y ⊆ X, with fX ⊆ Y ∩ gX rather than completeness of whole space X, • Theorem 2.23 is improved by replacing d-self-closedness or (g, d)-self- closedness of R instead of regularity of the whole space, • some examples are addopted to demonstrate the realized improvement in the results proved in this article. 3. Main Results Now, we are equipped to prove our main result as follows: Theorem 3.1. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R. Assume that the conditions (f), (g), (h), (i) and together with the following conditions hold: (m) there exists ϕ ∈ Φ such that d(fu, fv) ≤ ϕ(Mf (gu, gv)) (for all u, v ∈ X with (gu, gv) ∈ R); (k′) (k1) Y ⊆ g(X); (k′2) either f is (g, R)-continuous, or (f is continuous and g is bi- continuous), or R|Y is d-self-closed; or, alternatively (l′) (l′1) (f, g) is R-compatible; (l′2) g is R-continuous and (either f is R-continuous or (X, d) is (g, R)- regular); or, alternatively (l′′2 ) g is continuous and (either f is R-continuous or R is (g, d)-self- closed). Then (f, g) has a coincidence point. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 70 Relation-theoretic metrical coincidence and common fixed point theorems Proof. Let w0 ∈ X such that (gw0, fw0) ∈ R. Construct a Picard Jungck sequence {gwn}, with the initial point w0, i.e., (3.1) g(wn+1) = f(wn), for all n ∈ N0. Also as (gw0, fw0) ∈ R and R is (f, g)-closed, we have (fw0, fw1), (fw2, fw3), · · · , (fwn, fwn+1), · · · ∈ R. Thus, (3.2) (gwn, gwn+1) ∈ R, for all n ∈ N0, therefore {gwn} is R-preserving. From condition (m), we have (for all n ∈ N) (3.3) d(gwn, gwn+1) = d(fwn−1, fwn) ≤ ϕ(Mf (gwn−1, gwn)) where, Mf (gwn−1, gwn) ≤ max { d(gwn−1, gwn), d(gwn−1, fwn−1), d(gwn, fwn), 1 2 [ d(gwn, fwn−1) + d(gwn−1, fwn) ]} on using (3.1) and tringular inequality, we have (for all n ∈ N) (3.4) Mf(gwn−1, gwn) ≤ max { d(gwn−1, gwn), d(gwn, gwn+1) } . On using (3.3), (3.4) and the property (Φ1), we obtain (for all n ∈ N) (3.5) d(gwn, gwn+1) ≤ ϕ ( max { d(gwn−1, gwn), d(gwn, gwn+1) }) . Now, we show that the sequence {gwn} is Cauchy in (X, d). In case gwn0 = gwn0+1 for some n0 ∈ N0, then the result is follows. Otherwise, gwn 6= gwn+1 for all n ∈ N0. Suppose that d(gwn1−1, gwn1) ≤ d(gwn1, gwn1+1), for some n1 ∈ N. On using (3.5) and Lemma 2.21, we get d(gwn1, gwn1+1) ≤ ϕ(d(gwn1 , gwn1+1)) < d(gwn1, gwn1+1), which is a contradiction. Thus d(gwn, gwn+1) < d(gwn−1, gwn) (for all n ∈ N), so that d(gwn, gwn+1) ≤ ϕ(d(gwn−1, gwn)), for all n ∈ N. Employing induction on n and the property (Φ1), we get d(gwn, gwn+1) ≤ ϕn(d(gw0, gw1)), for all n ∈ N0. Now, for all m, n ∈ N0 with m ≥ n, we have d(gwn, gwm) ≤ d(gwn, gwn+1) + d(gwn+1, gwn+2) + · · · + d(gwm−1, gwm) ≤ ϕn(d(gw0, gw1)) + ϕn+1(d(gw0, gw1)) + · · · + ϕm−1(d(gw0, gw1)) = m−1 ∑ k=n ϕk(d(gw0, gw1)) ≤ ∑ k≥n ϕk(d(gw0, gw1)) → 0 as n → ∞. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 71 M. Ahmadullah, M. Imdad and M. Arif Therefore, {gwn} is R-preserving Cauchy sequence in X. As {gwn} ⊆ g(X) and {gwn} ⊆ Y ⊆ g(X) (due to (3.1) and (k1)), therefore {gwn} is R- preserving Cauchy sequence in Y. Since (Y, d) is R-complete, there exists y ∈ Y such that gwn d−→ y. As Y ⊆ g(X), there exists x ∈ X such that (3.6) lim n→∞ gwn = y = gx. Since f is (g, R)-continuous, and on using (3.1) and (3.6), we have (3.7) lim n→∞ gwn+1 = lim n→∞ fwn = fx. Due to uniqueness of the limit, we have fx = gx. Hence x is a coincidence point of (f, g). Next, we assume that f is continuous and g is bi-continuous. Then on using (3.1) and (3.6), we get fx = fg−1(gx) = fg−1( lim n→∞ gwn) = lim n→∞ fg−1(gwn) = lim n→∞ fwn = gx. Hence x is a coincidence point of (f, g). Finally, if R|Y is d-self-closed, then for any R-preserving sequence {gwn} in Y with gwn d−→ gx, there is a subsequence {gwnk} of {gwn} such that [gwnk , gx] ∈ R|Y ⊆ R, for all k ∈ N0. Set δ := d(fx, gx) ≥ 0. Suppose on contrary that δ > 0. On using condition (m), Proposition 2.22 and [gwnk , gx] ∈ R, for all k ∈ N0, we have (3.8) d(gwnk+1, fx) = d(fwnk , fx) ≤ ϕ(Mf (gwnk , gx)), where, Mf (gwnk , gx) = max { d(gwnk , gx), d(gwnk , gwnk+1), d(gx, fx), 1 2 [d(gwnk , fx) + d(gx, gwnk+1)] } . If Mf(gwnk , gx) = d(gx, fx) = δ, then (3.8) reduces to d(gwnk+1, fx) ≤ ϕ(δ), which on making k → ∞, gives arise δ ≤ ϕ(δ), which is a contradiction. Otherwise, if Mf (gwnk, gx) = max { d(gwnk , gx), d(gwnk , gwnk+1), 1 2 [d(gwnk , fx) + d(gx, gwnk+1)] } , then due to the fact that gwn d−→ gx, there exists a positive integer N = N(δ) such that Mf (gwnk , gx) ≤ 4 5 δ, for all k ≥ N. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 72 Relation-theoretic metrical coincidence and common fixed point theorems As ϕ is increasing, we have (3.9) ϕ(Mf (gwnk, gx)) ≤ ϕ( 4 5 δ), for all k ≥ N. On using (3.8) and (3.9), we get d(gwnk+1, fx) = d(fwnk , fx) ≤ ϕ( 4 5 δ), for all k ≥ N. Letting k → ∞ and using Lemma 2.21, we get δ ≤ ϕ(4 5 δ) < 4 5 δ < δ, which is again a contradiction. Hence, δ = 0, so that d(fx, gx) = δ = 0 ⇒ fx = gx. Thus, x is a coincidence point of (f, g). Alternatively, we suppose that (l′) holds. Firstly, assume that (l′2) holds. As {gwn} ⊂ f(X) ⊆ Y (in view (3.1)) we notice that {gwn} is R-preserving Cauchy sequence in Y. Since Y is R-complete, there exists y ∈ Y such that (3.10) lim n→∞ gwn = y and lim n→∞ fwn = y. As {fwn} and {gwn} are R-preserving sequences (due to (3.1) and (3.2)), utilizing the condition (l′1) and (3.10), we obtain (3.11) lim n→∞ d(gfwn, fgwn) = 0. Using (3.2), (3.10), and due to R-continuity of f and g, we have (3.12) lim n→∞ g(fwn) = g( lim n→∞ fwn) = gy, and (3.13) lim n→∞ f(gwn) = f( lim n→∞ gwn) = fy. On using (3.11)–(3.13) and continuity of d, we have fy = gy. Hence y is a coincidence point of (f, g). Next, assume that (X, d) is (g, R)-regular. As {gwn} is R-preserving and gwn d−→ y (due to (g, R)-regularity of (X, d)), we have [ggwn, gy] ∈ R, ∀ n ∈ N0. Set η := d(fy, gy) ≥ 0. Suppose on contrary that η > 0. On utilizing the condition (m), Proposition 2.22 and [ggwn, gy] ∈ R, for all n ∈ N0, we have (3.14) d(fgwn+1, fy) ≤ ϕ(Mf (ggwn+1, gy)), where, Mf(ggwn+1, gy) = max { d(ggwn+1, gy), d(ggwn+1, fgwn+1), d(gy, fy), 1 2 [d(ggwn+1, fy) + d(gy, fgwn+1)] } . c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 73 M. Ahmadullah, M. Imdad and M. Arif If Mf(ggwn, gy) = d(gy, fy) = η, then (3.14) yields |d(fgwn+1, gfwn+1) − d(gfwn+1, fy)| ≤ d(fgwn+1, fy) ≤ ϕ(η), on making n → ∞; using (3.10), (3.11), continuity of d and R-continuity of g, we get η ≤ ϕ(η), which is a contradiction. Therefore, η = 0, so that d(fy, gy) = η = 0 ⇒ fy = gy. Hence y is a coincidence point of (f, g). Otherwise, let Mf (ggwn+1, gy) = max { d(ggwn+1, gy), d(ggwn+1, fgwn+1), 1 2 [d(ggwn+1, fy) + d(gy, fgwn+1)] } . Now, on using triangular inequality, we have Mf (ggwn+1, gy) ≤ max { d(ggwn+1, gy), d(ggwn+1, ggwn+2) + d(ggwn+2, fgwn+1), 1 2 [d(ggwn+1, fy) + d(gy, gfwn+1) + d(gfwn+1, fgwn+1)] } . On making n → ∞, on using (3.1), (3.10), (3.11), continuity of d and R- continuity of g, we get lim n→∞ Mf(ggwn+1, gy) = 1 2 η. Since η > 0. By definition, there exists a positive integer N = N(η) such that Mf (ggwn+1, gy) ≤ 4 5 η, for all n ≥ N. As ϕ is increasing, we have ϕ(Mf (ggwn+1, gy)) ≤ ϕ( 4 5 η), for all n ≥ N, again (3.14) yields that ( for all n ≥ N) (3.15) |d(fgwn+1, gfwn+1) − d(gfwn+1, fy)| ≤ d(fgwn+1, fy) ≤ ϕ( 4 5 η). Hence, |d(fgwn+1, gfwn+1) − d(gfwn+1, fy)| ≤ ( 4 5 η), for all n ≥ N. Letting n → ∞, on using (3.10), (3.11), continuity of d and R-continuity of g, we get η ≤ ϕ(4 5 η) < 4 5 η < η, which is again a contradiction. Hence, η = 0, so that d(fy, gy) = η = 0 ⇒ fy = gy. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 74 Relation-theoretic metrical coincidence and common fixed point theorems Hence, y is a coincidence point of (f, g). Alternatively, suppose that (l′′2 ) holds. Firstly, we assume that f is R- continuous. Since continuity of g implies R-continuity of g. Thus on the similar lines of proof of (l′2) conclusion follows. Secondly, we assume that R is (g, d)-self-closed. As {gwn} is R-preserving and gwn d−→ y (due to (g, d)-self- closedness of R), there exists a subsequence {gwnk} of {gwn} such that [ggwnk, gy] ∈ R, ∀ k ∈ N0. Since gwn d−→ y, therefore gwnk d−→ y for any subsequence {gwnk} of {gwn}. Thus on the similar lines of above proof, we obtain (for all k ≥ N) (3.16) |d(fgwnk+1, gfwnk+1) − d(gfwnk+1, fy)| ≤ d(fgwnk+1, fy) ≤ ϕ( 4 5 η). Hence, |d(fgwnk+1, gfwnk+1) − d(gfwnk+1, fy)| ≤ ( 4 5 η), for all k ≥ N. Letting k → ∞, on using (3.10), (3.11), continuity of d and g, we get η ≤ ϕ(4 5 η) < 4 5 η < η, which is again a contradiction. Hence, η = 0, so that d(fy, gy) = η = 0 ⇒ fy = gy. Hence, y is a coincidence point of (f, g). This completes the proof. � On account taking Y = X in Theorem 3.1, we deduce a corollary which is sharpened version of Theorem 2.23 up to coincidence point in view of com- paratively weaker notions in the considerations of completeness, regularity and contraction condition. Corollary 3.2. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R. Suppose that the conditions (h), (i), (m) together with the following conditions hold: (n) (X, d) is R-complete; (o) f(X) ⊆ g(X); (p) g is onto with together the condition (k′2) [or, alternatively condition (l′)]. Then (f, g) has a coincidence point. In lieu of Remarks 2.20, Theorem 3.1 reduces to the following corollary. Corollary 3.3. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) endowed with a binary relation R and Y be a subspace of X. Assume the conditions (f), (g), (h), (i) and (k′) (or (l′)) together with the following condi- tion holds: c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 75 M. Ahmadullah, M. Imdad and M. Arif (q) there exists ϕ ∈ Φ such that (for all u, v ∈ X with (gu, gv) ∈ R) d(fu, fv) ≤ ϕ ( max { d(gu, gv), 1 2 [d(gu, fu)+d(gv, fv)], 1 2 [d(gu, fv)+d(gv, fu)] ) . Then (f, g) has a coincidence point. Now, we establish the following results for the uniqueness of common fixed point (corresponding to Corollary 3.3): Theorem 3.4. In addition to the hypotheses of Corollary 3.3, suppose that the following condition holds: (r) f(X) is (g, Rs)-directed. Then (f, g) has a unique point of coincidence. Moreover, if (f, g) is weakly compatible, then (f, g) has a unique common fixed point. Proof. We prove the result in three steps. Step 1: By Corollary 3.3, C(f, g) is non-empty. If C(f, g) is singleton, then there is nothing to prove. Otherwise, to substantiate the proof, take two arbi- trary elements u, v in C(f, g), so that fu = gu = x and fv = gv = y. Now, we are required to show that x = y. Since x, y ∈ fX and fX is (g, Rs)- directed, there exists u0 ∈ X such that [x, gu0] ∈ R and [y, gu0] ∈ R. Now, we construct a sequence {gun} corresponding to u0, so that gun+1 = fun for all n ∈ N0. We claim that lim n→∞ d(x, gun) = 0. If d(x, gun0) = 0, for some n0 ∈ N0, then there is nothing to prove. Otherwise, d(x, gun) > 0, for all n ∈ N0. As [x, gun] ∈ R, for all n ∈ N0 (due to the fact that (f, g)-closedness of R and [x, gu0] ∈ R), by Proposition 2.22 and hypothesis (q), we get (3.17) d(x, gun+1) = d(fu, fun) ≤ ϕ(Nf (gu, gun)), where, Nf (gu, gun) = max { d(gu, gun), 1 2 [d(gu, fu) + d(gun, fun)], 1 2 [d(gu, fun) + d(gun, fu) ]} ≤ max { d(gu, gun), 1 2 [d(gun, gu) + d(fun, gu)], 1 2 [d(gu, fun) +d(gun, fu)] } ≤ max { d(gu, gun), 1 2 [d(gun, gu) + d(fun, gu)] } ≤ max { d(gu, gun), d(gu, fun) } = max { d(gu, gun), d(gu, gun+1) } , c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 76 Relation-theoretic metrical coincidence and common fixed point theorems on using this and property (Φ1) (3.17) yields (for all n ∈ N0) d(gu, gun+1) ≤ ϕ ( max { d(gu, gun), d(gu, gun+1) }) = ϕ ( d(gu, gun) ) , otherwise, we get a contradiction. So, by induction on n, we get d(gu, gun) ≤ ϕn(d(gu, gu0)), for all n ∈ N0, which on making n → ∞ and using the property (Φ2), we get (3.18) lim n→∞ d(gu, gun) = 0. Similarly, we can obtain (3.19) lim n→∞ d(gv, gun) = 0. Using (3.18) and (3.19) we have d(x, y) ≤ d(gu, gun) + d(gun, gv) → 0, as n → ∞ ⇒ x = y i.e., (f, g) has a unique point of coincidence . Step 2: Now, we claim that the pair (f, g) has a common fixed point, let x ∈ C(f, g), i.e., fx = gx. Due to weakly compatibility of the pair (f, g), we have (3.20) f(gx) = g(fx) = g(gx). Put gx = y. Then from (3.20), fy = gy. Hence y is also a coincidence point of f and g. In view of Step 1, we have y = gx = gy = fy, so that y is a common fixed point (f, g). Step 3: To prove the uniqueness of common fixed point of (f, g), let us assume that w is another common fixed point of (f, g). Then w ∈ C(f, g), by Step 1, w = gw = gy = y. Hence (f, g) has a unique common fixed point. � Remark 3.5. In view of Theorem 3.4, we have used comparatively more nat- ural condition “(g, Rs)-directedness of f(X)” instead of “(g, Rs)-directedness of C(f, g)” which is too restrictive. Our proof carry on even if we take “C(f, g) is (g, Rs)-directed”. Since point of coincidence implies that coinci- dence point due to weakly compatible of (f, g), as in our Theorem 3.4 we want to find unique common fixed point of f and g which is the point in C(f, g). Theorem 3.6. In addition to the hypotheses of Theorem 3.1, assume the con- dition (r) together with the following condition holds: (s) one of f and g is one to one . Then (f, g) has a unique coincidence point. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 77 M. Ahmadullah, M. Imdad and M. Arif Proof. Due to Theorem 3.1, C(f, g) 6= ∅. Let u, v ∈ C(f, g), and hence in similar lines of the proof of Theorem 3.4, we have gu = fu = fv = gv. Since either f or g is one-one, we have u = v. � Notice that Theorem 3.6 is a natural improved version of Theorem 4 due to Alam and Imdad [5]. Theorem 3.7. In addition to the hypotheses of Theorem 3.1, assume the fol- lowing condition holds: (t) R|fX is complete. Then (f, g) has a unique point of coincidence. Moreover, if (f, g) is weakly compatible, then (f, g) has a unique common fixed point. Proof. From Theorem 3.1, we have C(f, g) 6= ∅. If C(f, g) is singleton, then proof is over. Otherwise, choose any two elements x 6= y in C(f, g), so that fx = gx = x and fy = gy = y. As R|fX is complete, [x, y] ∈ R. Using Proposition 2.22 and condition (m), we get d(x, y) = d(fx, fy) ≤ ϕ ( max { d(gx, gy), d(gx, fx), d(gy, fy), 1 2 [d(gx, fy) + d(gy, fx)] }) = ϕ ( d(gx, gy) ) < d(gx, gy) = d(x, y), which is a contradiction, hence d(x, y) = 0, therefore x = y. Thus (f, g) has a unique point of coincidence. Thus the remaining part of the proof can be obtained from Theorem 3.4. � Remark 3.8. Indeed, Theorem 3.7 is more general as compared to Corollary 5.1 of Berzig [9] and Corollary 3 due to Alam and Imdad [5]. In regard of Remark 3.5, on considering symmetric closure S of any binary relation R in Theorem 3.4, we obtain the following sharpened version of The- orem 2.23. Corollary 3.9. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) endowed with symmetric closure S of any arbitrary binary relation de- fined on X and Y be a subspace of X. Assume that the conditions (b), (c), (e) and (g) together with the following conditions hold: (u) (Y, d) is S-complete subspace of X; (v) (v1) Y ⊆ g(X); c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 78 Relation-theoretic metrical coincidence and common fixed point theorems (v2) either f is (g, S)-continuous or (f is continuous and g is bi- continuous) or S|Y is d-self-closed; or, alternatively (w) (w1) (f, g) is S-compatible; (w2) g is S-continuous and (either f is S-continuous or (X, d) is (g, S)- regular); or, alternatively (w′2) g is continuous and (either f is S-continuous or S is (g, d)-self- closed). Then (f, g) has a coincidence point. Moreover, if C(f, g) is (g, Rs)-directed and (f, g) is weakly compatible, then (f, g) has a unique common fixed point. Notice that the hypotheses ‘S is (f, g)-closed’ is equivalent to f is a ‘g- comparative’ and ‘S|Y is d-self-closed’ is more natural ‘the regular property of (Y, d, S)’. Further ‘S is (g, d)-self-closed’ is more natural the ‘S is d-self-closed’. 4. Consequences As consequences of our former proved results, we deduce several well known results of the existing literature. On the setting ϕ(t) = kt, with k ∈ [0, 1), we obtain the following corollaries which are immediate consequences of Theorem 3.4. Corollary 4.1. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R and Y a subspace of X. Suppose that the conditions (f), (g), (h), (i) and (k′) (or (l′)) together with the following con- dition holds: (q1) there exists k ∈ [0, 1] such that ( for all u, v ∈ X with (gu, gv) ∈ R) d(fu, fv) ≤ k(max { d(gu, gv), 1 2 [d(gu, fu) + d(gv, fv)], 1 2 [d(gu, fv) + d(gv, fu)] }) . Then (f, g) has a coincidence point. Moreover, if f(X) is (g, Rs)-directed and (f, g) is weakly compatible, then (f, g) has a unique common fixed point. Remark 4.2. Corollary 4.1 is a sharpened version of Corollary 5.10 of Berzig [9] and Corollary 3.3 (corresponding to condition (20)) due to Ahmadullah et al. [2]. Corollary 4.3. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R. Suppose that the conditions (f), (g), (h), (i) and (k′) (or (l′)) together with the following condition holds: (q2) there exist a, b, c ≥ 0 with a+2b+2c < 1 such that (for all u, v ∈ X with (gu, gv) ∈ R) d(fu, fv) ≤ ad(gu, gv) + b[d(gu, fu) + d(gv, fv)] + c[d(gu, fv) + d(gv, fu)]. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 79 M. Ahmadullah, M. Imdad and M. Arif Then (f, g) has a coincidence point. Moreover, if f(X) is (g, Rs)-directed and (f, g) is weakly compatible, then (f, g) has a unique common fixed point. Remark 4.4. Corollary 4.3 remains a sharpened version of Corollary 5.11 due to Berzig [9] and Corollary 3.3, (corresponding to condition (22)) in view of Ahmadullah et al. [2]. Remark 4.5. If b = 0 and c = 0 in Corollary 4.3, then we deduces Theorem 2.24 (see Alam and Imdad [5]). Corollary 4.6. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped a binary relation R and Y a subspace of X. Assume that the conditions (f), (g), (h), (i) and (k′) (or (l′)) together with the following condi- tion holds: (q3) there exists k ∈ [0, 1/2) such that (for all u, v ∈ X with (gu, gv) ∈ R) d(fu, fv) ≤ k[d(gu, fu) + d(gv, fv)]. Then (f, g) has a coincidence point. Moreover, if f(X) is (g, Rs)-directed and (f, g) is weakly compatible, then (f, g) has a unique common fixed point. Remark 4.7. Corollary 4.6 remains a improved version of Corollary 5.13 es- tablished in Berzig [9] and Corollary 3.3 (corresponding to condition (18)) in Ahmadullah et al. [2]. Corollary 4.8. Let (f, g) be a pair of self-mappings defined on a metric space (X, d) equipped with a binary relation R and Y a subspace of X. Assume that the conditions (f), (g), (h), (i) and (k′) (or (l′)) togetrher with the following condition holds: (q4) there exists k ∈ [0, 1/2) such that (for all u, v ∈ X with (gu, gv) ∈ R) d(fu, fv) ≤ k [ d(gu, fv) + d(gv, fu) ] . Then (f, g) has a coincidence point. Moreover, if f(X) is (g, Rs)-directed and (f, g) is weakly compatible, then (f, g) has a unique common fixed point. Remark 4.9. Corollary 4.8 is an improved version of Corollary 5.14 of Berzig [9] and Corollary 3.3 (corresponding to condition (19)) due to Ahmadullah et al. [2]. Remark 4.10. Under the consideration g = I (identity mapping on X), The- orems 3.1 and 3.4 deduce the fixed point results of Ahmadullah et al. [3, Theorem 2.1 and 2.5]. Remark 4.11. On setting g = I (identity mapping on X), in Corollaries 3.2-4.8, we deduce the fixed point results which are the sharpened version of several results in the existing literature. Under the universal relation (i.e., R = X × X), Theorems 3.4 and 3.7 unify to the following lone corollary: c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 80 Relation-theoretic metrical coincidence and common fixed point theorems Corollary 4.12. Let (X, d) be a metric space and (f, g) a pair of self-mappings on X. Suppose that the following conditions hold: (A) there exists Y ⊆ X, f(X) ⊆ Y ⊆ g(X) such that (Y, d) is complete; (B) there exists ϕ ∈ Φ such that (for all u, v ∈ X) d(fu, fv) ≤ ϕ ( max { d(gu, gv), 1 2 [d(gu, fu) + d(gv, fv)], 1 2 [d(gu, fv) + d(gv, fu)] }) . Then (f, g) has a unique common fixed point. 5. Illustrative Examples In this section, we furnish some examples to demonstrate the realized im- provement of our proved results. Example 5.1. Let (X, d) be a metric space, where X = (−2, 4) and d(x, y) = |x − y|, ∀ x y ∈ X. Now, define a binary relation R = {(x, y) ∈ X2 | x ≥ y, x, y ≥ 0} ∪ {(x, y) ∈ X2 | x ≤ y, x, y ≤ 0}, an increasing mapping ϕ : [0, ∞) → [0, ∞) by ϕ(s) = 1 3 s and two self-mappings f, g : X → X by f(x) = 0, ∀ x ∈ (−2, 4) and g(x) = { x 3 , x ∈ (−2, 3]; 1, x ∈ (3, 4). Let Y = [−1 2 , 1), so that f(X) = {0} ⊂ Y ⊂ gX = (−2 3 , 1] and Y is R- complete but X is not R-complete. Indeed, R is (f, g)-closed, f and g are R-continuous and (f, g) is R-compatible. By straightforward calculations, one can easily verify hypothesis (m) of Theorems 3.1 thus in all by Theorem 3.1 we obtain, (f, g) has a coincidence point (Observe that, C(f, g) = {0}). More- over, as fX is (g, Rs)-directed, R|fX is complete and (f, g) commute at their coincidence point i.e., x = 0 therefore, all the hypotheses of Theorems 3.4 and 3.7 are satisfied, ensuring the uniqueness of the common fixed point. Notice that, x = 0 is the only common fixed point of (f, g). With a view to show the genuineness of our results, notice that R is not symmetric and R can not be a symmetric closure of any binary relation. Also (X, d) is not complete and even not R-complete which shows that Theorems 3.1, 3.4 and 3.7 are applicable to the present example, while Theorem 2.23 and even Corollary 3.2 are not, which substantiates the utility of Theorems 3.1, 3.4 and 3.7. Example 5.2. Let X = [0, 4) with usual metric d and R = {(0, 0), (0, 1), (1, 0), (1, 1), (1, 2), (2, 3)} be a binary relation whose symmetric closure S = {(0, 0), (0, 1), (1, 0), (2, 3), (1, 1), (1, 2), (2, 1), (3, 2)} and (f, g) a pair of self-mappings on X defined by f(x) = { 0, x ∈ [0, 4) ∩ Q; 1, x ∈ [0, 4) ∩ Qc, and g(x) = { x, x ∈ {0, 1, 2}; 3, otherwise. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 81 M. Ahmadullah, M. Imdad and M. Arif Let Y = {0, 1} which is R-complete and fX = {0, 1} ⊆ Y ⊆ gX = {0, 1, 2, 3}. Define an increasing function ϕ : [0, ∞) → [0, ∞) by ϕ(s) = 5 6 s. Clearly, ϕ ∈ Φ and both f and g are not continuous. Also, R is (f, g)-closed. Take any R-preserving sequence {xn} in Y, i.e., (xn, xn+1) ∈ R|Y , for all n ∈ N with xn d−→ x. Here, one can notice that if (xn, xn+1) ∈ R|Y , for all n ∈ N then there exists N ∈ N such that xn = x ∈ {0, 1}, for all n ≥ N. So, we choose a subsequence {xnk} of the sequence {xn} such that xnk = x, for all k ∈ N, which amounts to saying that [xnk , x] ∈ R|Y , for all k ∈ N. Therefore, R|Y is d-self-closed. Now, to substantiate the contraction condition (m) of Theorems 3.1. For this, we need to verify for (gx, gy) ∈ {(2, 3)}, otherwise, d(fx, fy) = 0. If (gx, gy) ∈ {(2, 3)} ⇒ x = 2, y ∈ [0, 4) − {0, 1, 2}, then there are two cases arises: Case (1): If x = 2, y ∈ ([0, 4) − {0, 1, 2}) ∩ Q, then condition (m) is obvious. Case (2): If x = 2, y ∈ ([0, 4) − {0, 1, 2, 3}) ∩ Qc, then we have d(f2, fy) = 1 ≤ ϕ(max{d(g2, gy), d(g2, f2), d(gy, fy), 1 2 [d(g2, fy) + d(gy, f2)]}) = ϕ(max{d(2, 3), d(2, 0), d(3, 1), 1 2 [d(2, 1) + d(3, 0)]} = ϕ(2). Thus all the conditions of Theorem 3.1 are satisfied, hence (f, g) has a co- incidence point (namely C(f, g) = {0}). Also fX is (g, Rs)-directed, (f, g) commutes at their coincidence point i.e., at x = 0 and condition (m) of Theo- rem 3.4 holds. Therefore all the hypotheses of Theorem 3.4 are satisfied. Notice that, x = 0 is the only common fixed point of (f, g). Now, since (gx, gy) = (2, 3) ∈ R, clearly x = 2, we choose y = √ 2 but 1 = d(f2, f √ 2) ≤ αd(g2, g √ 2) = α, which shows that contraction condition of Theorem 2.24 (due to Alam and Imdad [5]) is not satisfied. Further, Theorem 2.23 is not applicable to the present example as underlying metric space (X, d) is not complete and R is not symmetric closure of any binary relation. Thus, our results are an improvement over Theorem 2.23 (due to Berzig [9]) and Theorem 2.24 (Alam and Imdad [5]). Acknowledgements. All the authors are thankful to an anonymous referee for his/her valuable comments and suggestions. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 82 Relation-theoretic metrical coincidence and common fixed point theorems References [1] M. Ahmadullah, J. Ali and M. Imdad, Unified relation-theoretic metrical fixed point the- orems under an implicit contractive condition with an application, Fixed Point Theory Appl. 2016:42 (2016), 1–15 . [2] M. Ahmadullah, M. Imdad and M. Arif, Common fixed point theorems under an im- plicit contractive condition on metric spaces endowed with an arbitrary binary relation, preprint (arXiv:1701.03154 (2017)). [3] M. Ahmadullah, M. Imdad and R. Gubran, Relation-theoretic metrical fixed point the- orems under nonlinear contractions, Fixed Point Theory, to apper (arXiv:1611.04136v1 (2016)). [4] A. Alam and M. Imdad, Relation-theoretic contraction principle, J. Fixed Point Theory Appl. 17, no. 4 (2015), 693–702 . [5] A. Alam and M. Imdad, Relation-theoretic metrical coincidence theorems, Filomat 31, no. 14 (2017), 4421–4439. [6] A. Alam and M. Imdad, Nonlinear contractions in metric spaces under locally T-transitive binary relations, Fixed Point Theory, to appear (arXiv:1512.00348v1 (2015)). [7] M. Altman, An integral test for series and generalized contractions, Am. Math. Mon. 82 (1975), 827–829. [8] S. Banach, Sur les opérations dans les ensembles abstraits et leur application aux équations intégrales, Fund. Math. 3 (1922), 133–181. [9] M. Berzig, Coincidence and common fixed point results on metric spaces endowed with arbitrary binary relation and applications, J. Fixed Point Theory Appl. 12 (2012), 221– 238. [10] H. Ben-El-Mechaiekh, The Ran-Reurings fixed point theorem without partial order: A simple proof, J. Fixed Point Theory Appl. 16 (2015), 373–383. [11] L.B. Ćirić, A generalization of Banach’s contraction principle, Proc. Am. Math. Soc. 45 (1974), 267–273. [12] S. Ghods, M. E. Gordji, M. Ghods and M. Hadian, Comment on “Coupled fixed point theorems for nonlinear contractions in partially ordered metric spaces” [Lakshmikantham and Ćirić, Nonlinear Anal. TMA 70 (2009) 4341-4349], J. Comput. Anal. 14, no. 5 (2012), 958–966. [13] R. H. Haghi, Sh. Rezapour and N. Shahzad, Some fixed point generalizations are not real generalization, Nonlinear Anal. 74 (2011), 1799–1803. [14] J. Jachymski, The contraction principle for mappings on a metric space with a graph, Proc. Am. Math. Soc. 136 (2008), 1359–1373. [15] G. Jungck, Commutings maps and fixed points, Am. Math. Mon. 83, no. 4 (1976), 261–263. [16] G. Jungck, Common fixed points for noncontinuous nonselfs maps on non-metric space, Far East J. Math. Sci. 4 (1996), 199–255. [17] G. Jungck, Compatible mappings and common fixed points, Int. J. Math. Sci. 9, no. 4 (1986), 771–779. [18] S. Lipschutz, Schaum’s outlines of theory and problems of set theory and related topics, McGraw-Hill, New York (1964). [19] R. D. Maddux, Relation algebras, Studies in Logic and the Foundations of Mathematics, 150, Elsevier B. V., Amsterdam (2006). [20] J. J. Nieto and R. Rodŕıguez-López, Contractive mapping theorems in partially ordered sets and applications to ordinary differential equations, Order 22, no. 3 (2005), 223–239. [21] A. C. M. Ran and M. C. B. Reurings, A fixed point theorem in partially ordered sets and some applications to matrix equations, Proc. Am. Math. Soc. 132, no. 5 (2004), 1435–1443. [22] B. Samet and M. Turinici, Fixed point theorems on a metric space endowed with an arbitrary binary relation and applications, Commun. Math. Anal. 13 (2012), 82–97. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 83 M. Ahmadullah, M. Imdad and M. Arif [23] M. Turinici, Fixed points for monotone iteratively local contractions, Dem. Math. 19, no. 1 (1986), 171–180. [24] M. Turinici, Linear contractions in product ordered metric spaces, Ann. Univ. Ferrara. 59 (2013), 187–198. [25] M. Turinici, Ran-Reurings fixed point results in ordered metric spaces, Libertas Math. 31 (2011), 49–55. [26] M. Turinici, Nieto-Lopez theorems in ordered metric spaces, Math. Student 81, no. 1-4 (2012), 219–229. [27] S. Sessa, On a weak commutativity condition of mappings in fixed point considerations, Pub. Inst. Math. Soc. 32 (1982),149–153. [28] K. P. R. Sastry and I. S. R. Krishna Murthy, Common fixed points of two partially commuting tangential selfmaps on a metric space, J. Math. Anal. Appl. 250, no. 2 (2000), 731–734. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 84