We introduce the notion of graded-edge Rakotch contraction mappings on a generalized metric space with a weighted graph. We prove that the successive iterate of every starting point converge uniformly to a common limit, which is the unique fixed point. We derive an explicit iteration-complexity bound and, under a relaxed triangle inequality, a geometric rate whose constant is the graph invariant. Finally, we illustrate the theory on an explicit real graded contraction that realizes the interior optimum, on a quadratic integral equation, and on finite weighted digraphs.
Introduction
Banach’s contraction principle [2] has led to a vast amount of literature, of which we mention only the parts directly relevant here. Rakotch [9] replaced the constant Lipschitz factor by a decreasing comparison function φ subject to φ(t)<1 for t>0, and thereby obtained a strictly larger class of mappings retaining the conclusion of Banach’s theorem; the point of the relaxation is precisely that φ(0) is allowed to equal 1, so that a Rakotch contraction need not be Lipschitz with a constant smaller than one. A second line of development concerns spaces whose distance function satisfies only a relaxed triangle inequality d(x,z)≤M(d(x,y)+d(y,z)) for a constant M≥1. Such spaces, the b-metric spaces of Bakhtin [1] and Czerwik [6], form a proper generalization of metric spaces, the case M=1, and support their own extensive fixed point theory. A standard instance is the snowflake d(x,y)=∣x−y∣p with p>1 on the real line, for which M=2p−1 and the ordinary triangle inequality fails.
Motivated by such examples, Reich and Zaslavski [11] introduced generalized metric spaces, in which the inequality d(x,z)≤Md(x,y)+Md(y,z) is taken as an axiom, and proved a Rakotch-type fixed point theorem. In the recent paper [12] they went further in two directions: First, they replaced the relaxed triangle inequality by a purely qualitative ε–δ condition, so that no numerical constant is available at all. Second, they endowed the space with a graph G and required the contractive estimate to hold only along the edges of G, the two points of an arbitrary pair being connected instead by a G-chain of length at most a fixed integer n0. This last device goes back to Jachymski [7], whose theorem unified the classical contraction principle with its ordered-space variants due to Ran and Reurings [10] and Nieto and Rodríguez-López [8]; Beg et al. [3] see also Bojor [5] and Beg and Butt [4].
In [12] the graph enters twice: once through the edge-preservation requirement, and once through the connectivity hypothesis (P3). The contractive estimate itself, however, is governed by a single scalar comparison function φ, so that the graph contributes no quantitative information beyond the combinatorial bound n0. We propose to let the graph carry the contraction data as well. Concretely, we attach to each edge a grade w(e)∈(0,1], we let Eα={e∈E(G):w(e)≥α} denote the associated level cuts, and we ask that the mapping contract with comparison function φα on Eα, where α↦φα is pointwise non-increasing.
Substituting φ:=φα0 with α0=infew(e) restores the ungraded hypothesis, so that the qualitative conclusions cannot change. What makes the refinement substantive is that the two roles of the graph now move in opposite directions. Raising α improves the contraction factor but deletes edges, so that the chain-length invariant N(α):=supx,y∈KρGα(x,y) increases. Every convergence estimate we prove therefore depends on α through two competing terms and must be optimized over the threshold. In Example 6.11 we exhibit an explicit space, mapping and grading for which the optimal threshold is interior, and for which the ungraded choice α0 is strictly worse than the graded optimum by a factor that persists as the tolerance tends to zero. We believe this example and not the definition itself as justification of the framework.
The second observation is related to the origin of the grading. It does not need to be provided as additional information: it is already contained in any edge-contractive mapping. Defining c(x,y) to be the supremum of the orbital ratios d(Tn+1x,Tn+1y)/d(Tnx,Tny) and putting w:=1−c produces a grade which automatically satisfies the monotonicity axiom, and which normalizes the comparison family to φα≡1−α (Proposition 3.4). The graded theory is therefore not an additional hypothesis imposed on T, but a decomposition of the information T already carries.
Our contributions are the following.
We introduce graded-edge Rakotch contractions (Definitions 3.1 and 3.2) over the qualitative framework of [12], retaining the possible asymmetry of d and treating G as a directed graph throughout, and we show that the grading always exists canonically (Proposition 3.4).
We make the chaining step of [12] constructive by exhibiting an explicit modulus ηm obtained by iterating the modulus of the ε–δ triangle condition (Lemma 4.3).
We prove a global uniform convergence theorem (Theorem 5.2) which contains [12 Theorem 2] as the case w≡1, and we close a gap in [12 Theorem 2], whose statement asserts the convergence of the Picard iterates but not the fixed point property of the limit (Theorem 5.5).
We derive an explicit iteration-complexity bound (Theorem 6.1) and, under a relaxed triangle inequality, a geometric rate whose constant is the graph invariant N(α)MN(α)−1 (Theorem 6.4); we optimize over α (Corollary 6.10) and settle the resulting trade-off explicitly (Example 6.11).
We illustrate the theory on an explicit real graded contraction that realizes the interior optimum, on a quadratic integral equation, and on finite weighted digraphs (Section 7).
Preliminaries
The space
Throughout, X is a nonempty set and d:X×X→[0,∞) satisfies
d(x,y)=0⟺x=y(x,y∈X).(1)
Following [12] we do not assume that d is symmetric, nor that it satisfies a triangle inequality; instead we impose the following two conditions.
Assumption 2.1 (A1). For each ε>0 there exists δ>0 such that d(x,y)≤δ implies d(y,x)≤ε.
Assumption 2.2 (A2). For each ε>0 there exists δ>0 such that d(x,y)≤δ and d(y,z)≤δ imply d(x,z)≤ε.
For ε>0 we fix once and for all a choice of δ1(ε) for (A1) and δ2(ε) for (A2). Without loss of generality, we may assume that
δi is non-decreasing and δi(ε)≤ε(i=1,2,ε>0).(2)
We refer to δ1 and δ2 as the symmetry modulus and the triangle modulus of (X,d).
Setting
U(ε):={(x,y)∈X×X:d(x,y)≤ε,d(y,x)≤ε},ε>0,(3)
one checks as in [12] that {U(ε)}ε>0 is a base of a uniformity on X, metrizable by a metric d1. We assume throughout that the metric space (X,d1) is complete. By construction, for every ρ>0 there is ε>0 with U(ε)⊆{(x,y):d1(x,y)≤ρ}, and conversely for every ε>0 there is ρ>0 with {(x,y):d1(x,y)≤ρ}⊆U(ε); in particular
d1(xn,x)→0⟹d(xn,x)→0 and d(x,xn)→0.(4)
A set E⊆X is bounded if sup{d(x,y):x,y∈E}<∞.
Remark 2.3. In several places below we shall strengthen (A2) to the quantitative condition
d(x,z)≤M(d(x,y)+d(y,z))(x,y,z∈X),(5)
which is the notion of a generalized metric space of [11]; when d is symmetric this is exactly a b-metric space with constant M in the sense of [1, 6]. Taking y=x in (5) forces M≥1. Condition (5) implies (A2) with δ2(ε)=ε/(2M). We refer to the setting of (A1)–(A2) as the qualitative regime and to the setting of (5) as the quantitative regime. When M=1 and d is symmetric, (X,d) is an ordinary metric space.
Comparison functions
Let Φ denote the set of all decreasing functions φ:[0,∞)→[0,1] such that φ(t)<1 for every t>0.
Remark 2.4. The requirement φ(t)<1 is imposed only for t>0, so that φ(0)=1 is permitted. This is essential: a decreasing φ:[0,∞)→[0,1) would satisfy φ(t)≤φ(0)<1 for every t, so that the Rakotch condition would collapse to the Banach condition with constant φ(0) and no generality would be gained. The definition in [12 §1] has codomain [0,1], as here; the one in [12 §2] has codomain [0,1), which we take to be a slip of the pen.
Graded graphs
Let G be a directed graph with vertex set V(G)=X and edge set E(G)⊆X×X containing all loops. We write G for the symmetrization of G, that is, the undirected graph in which x and y are adjacent whenever (x,y)∈E(G) or (y,x)∈E(G).
Definition 2.5. A grade on G is a map w:E(G)→(0,1]. For α∈(0,1] the level cut of G at α is the directed graph Gα with V(Gα)=X and
Eα:={e∈E(G):w(e)≥α}.
The pair (G,w) is called a graded graph on X.
Clearly E1⊆Eα⊆Eβ whenever β≤α, so the level cuts form a decreasing family of graphs.
Given a nonempty K⊆X, a Gα-chain in K from x to y of length q is a finite sequence ξ0,…,ξq∈K with ξ0=x, ξq=y and, for each i∈{0,…,q−1},
(ξi,ξi+1)∈Eαor(ξi+1,ξi)∈Eα.
We denote by ρGα(x,y) the least length of such a chain (with ρGα(x,y)=∞ if none exists) and we set
N(α):=x,y∈KsupρGα(x,y)∈N∪{∞},(6)
the α-chain diameter of K. The function α↦N(α) is non-decreasing, and
A:={α∈(0,1]:N(α)<∞}
is a sub-interval of (0,1] containing 0 in its closure whenever it is nonempty. The connectivity hypothesis (P3) of [12 page 13] is precisely the statement N(α)<∞ for the ungraded graph, with n0=N(α).
Graded-edge Rakotch contractions
Throughout this section, (G,w) is a graded graph on X, K⊆X is nonempty and closed in (X,d1), and T:K→K.
Definition 3.1 (Grade-monotone edge preservation, (C1)). The mapping T is said to be grade-monotone edge preserving if for every (x,y)∈E(G) with x,y∈K,
(T(x),T(y))∈E(G)andw(T(x),T(y))≥w(x,y).
Condition (C1) strengthens the edge-preservation requirement of [12 (5)] by demanding that T never weakens a connection. It is what guarantees that an orbit which starts inside a level cut remains there.
Definition 3.2 (Graded Rakotch condition, (C2)). Let {φα}α∈(0,1]⊆Φ be a family such that
α≤β⇒φβ(t)≤φα(t)(t≥0),(7)
that is, α↦φα is pointwise non-increasing. The mapping T is a graded-edge Rakotch contraction with respect to (G,w) and {φα} if it satisfies (C1) and
d(T(x),T(y))≤φw(x,y)(d(x,y))d(x,y)for every (x,y)∈E(G) with x,y∈K.(8)
We set T0(x)=x for $x\in K.
Remark 3.3. If w≡1 and φ1=φ, then (C1) reduces to [12 (5)] and (8) reduces to [12 (6)]. If moreover E(G)=K×K, one recovers the hypotheses of [12 Theorem 1]. Finally, if each φα is constant, one recovers the graph-theoretic Banach contractions of [7].
The next result shows that Definition 3.2 does not in fact impose any new hypothesis on T beyond edge preservation and a mild uniformity: the grading may always be manufactured from the mapping itself, and when it is, the comparison family is normalized.
Proposition 3.4 (Canonical grading). Let T satisfy (x,y)∈E(G)⇒(T(x),T(y))∈E(G) for x,y∈K. For (x,y)∈E(G) with x,y∈K put
with the convention sup∅:=0, and assume c(x,y)<1 for every such pair. Then w:=1−c is a grade on E(G), and T is a graded-edge Rakotch contraction with respect to (G,w) and the family
φα≡1−α(α∈(0,1]).
Moreover 1−α is optimal, in the sense that no constant c′<1−α satisfies d(T(x),T(y))≤c′d(x,y) for every (x,y)∈Eα, whenever the value α is attained by w and the supremum in (9) is attained at n=0 for some such edge.
Proof. By hypothesis c(x,y)∈[0,1), so w(x,y)=1−c(x,y)∈(0,1] and w is a grade.
(C1). Edge preservation holds by assumption. Since the mapping T preserves edges, the pair (T(x),T(y)) lies in E(G) and, directly from (9),
the supremum on the right being taken over a subset of the index set defining c(x,y). Hence w(T(x),T(y))≥w(x,y).
(C2). Each φα≡1−α is a constant function with value in [0,1), hence lies in Φ, and α↦1−α is decreasing, so (7) holds. Let (x,y)∈E(G) with x,y∈K. If d(x,y)=0 then x=y and (8) is trivial. Otherwise the case n=0 of (9) gives
Optimality. Suppose (x,y)∈Eα with w(x,y)=α and that the supremum in (9) is attained at n=0. Then d(T(x),T(y))=c(x,y)d(x,y)=(1−α)d(x,y), so no smaller constant is admissible.
Remark 3.5. It is tempting to dismiss Definition 3.2 on the ground that φ:=φα0 with α0:=inf{w(e):e∈E(G)} satisfies (8) on all of E(G), so that every graded-edge Rakotch contraction is an ordinary one. This is correct, and it is exactly why the qualitative statements of Section 5 are not new in substance. The content of the grading lies in Section 6: the quantitative estimates involve φα and N(α) simultaneously, and passing to α0 minimizes the first factor while maximizing the second. Example 6.11 exhibits data for which the choice α0 is strictly worse than the optimal graded choice.
Auxiliary results
We first record the effect of (C1) along an orbit.
Lemma 4.1 (Invariance and monotonicity). Let T be a graded-edge Rakotch contraction, let α∈(0,1] and let x,y∈K satisfy (x,y)∈Eα. Then
(Tn(x),Tn(y))∈Eα(n≥0),
and the sequence n↦d(Tn(x),Tn(y)) is non-increasing. If moreover φα≡cα is constant, then
d(Tn(x),Tn(y))≤cαnd(x,y)(n≥0).(10)
Proof. The first assertion holds for n=0 by hypothesis. If (Tn(x),Tn(y))∈Eα, then (C1) gives (Tn+1(x),Tn+1(y))∈E(G) and
w(Tn+1(x),Tn+1(y))≥w(Tn(x),Tn(y))≥α,
so that (Tn+1(x),Tn+1(y))∈Eα; the claim follows by induction. For the second assertion, put an:=d(Tn(x),Tn(y)). Applying (8) to the edge (Tn(x),Tn(y))∈E(G) and using φw≤1 we obtain
an+1≤φw(Tn(x),Tn(y))(an)an≤an.
If φα≡cα, then w(Tn(x),Tn(y))≥α and (7) give an+1≤cαan, whence (10) by induction. □
The next lemma is the graded form of [12 Lemma 1]. It converts the multiplicative contraction (8) into an additive decrease that is uniform over all edges of a given level cut.
Lemma 4.2 (Graded uniform decrease). Let T be a graded-edge Rakotch contraction, let α∈(0,1], let ε>0, and let x,y∈K satisfy
(x,y)∈Eαandd(x,y)>ε.
Then
d(T(x),T(y))≤d(x,y)−ε(1−φα(ε)).
Proof. Write w:=w(x,y) and t:=d(x,y). Since (x,y)∈Eα we have w≥α, so that φw(t)≤φα(t) by (7). Since φα is decreasing and t>ε, we also have φα(t)≤φα(ε). Hence, by (8),
where the last inequality uses t>ε together with 1−φα(ε)≥0. Rearranging gives the assertion. □
In [12] the existence of a threshold ε0 with the property that short steps along a bounded chain produce a short displacement between its endpoints is deduced from (A2) but is not exhibited. The following lemma supplies it explicitly. Recall the moduli δ1,δ2 fixed in (2) and let δ2k denote the k-fold composition of δ2 with itself, with δ20:=id.
Let m≥1, let ε>0 and let ξ0,…,ξm∈X be such that for each i∈{0,…,m−1} at least one of the inequalities
d(ξi,ξi+1)≤ηm(ε),d(ξi+1,ξi)≤ηm(ε)
holds. Then d(ξ0,ξm)≤ε.
Proof. We argue by induction on m. Observe first that, by (2), each ηm is non-decreasing and ηm+1≤ηm≤η1≤id.
Case m=1. If d(ξ0,ξ1)≤η1(ε)≤ε there is nothing to prove. If instead d(ξ1,ξ0)≤η1(ε)≤δ1(ε), then Assumption 2.1 yields d(ξ0,ξ1)≤ε.
Inductive step. Let m≥1 and assume the assertion holds for chains of length m. Let ξ0,…,ξm+1 be a chain all of whose steps are bounded by ηm+1(ε), and put ε′:=δ2(ε). By (11),
ηm+1(ε)=η1(δ2m(ε))=η1(δ2m−1(ε′))=ηm(ε′).
Hence the first m steps of the chain are bounded by ηm(ε′), and the induction hypothesis, applied with ε′ in place of ε, gives
d(ξ0,ξm)≤ε′=δ2(ε).
The last step is bounded by ηm+1(ε)=ηm(ε′)≤η1(ε′), so the case m=1 applied to the pair (ξm,ξm+1) with ε′ in place of ε gives
d(ξm,ξm+1)≤ε′=δ2(ε).
Assumption 2.2 now yields d(ξ0,ξm+1)≤ε, which completes the induction.
Remark 4.4. In the quantitative regime (5) with d symmetric one may take δ1(ε)=ε and δ2(ε)=ε/(2M), whence
ηm(ε)=(2M)m−1ε(m≥1).
Thus the chaining modulus degrades geometrically in the chain length, at a rate governed by the constant M. This is the precise mechanism by which the graph invariant N(α) enters the estimates of § 7, and the reason why no analogue of those estimates is available in the purely qualitative regime.
Fixed point
Let K⊆X be a nonempty and closed in (X,d1),
D0:=sup{d(x,y):x,y∈K}<∞,(12)
T:K→K is a graded-edge Rakotch contraction with respect to a graded graph (G,w) and a family {φα} as in Definition 3.2, and A=∅, that is, N(α)<∞ for at least one α∈(0,1].
Assumption 5.1 (A3). ε0(ε,α) and k0(ε,α) are finite and positive.
Theorem 5.2 (Global uniform convergence). Let Assumption 5.1 hold and let α∈A. Then:
for everyε>0 and every integer n≥k0(ε,α),
x,y∈Ksupd(Tn(x),Tn(y))≤ε;(14)
there exists a point x∗∈K such that Tn(x)→x∗ in (X,d1) as n→∞, for every x∈K, and the convergence is uniform on K.
Proof. Write ε0:=ε0(ε,α) and k0:=k0(ε,α).
Step 1 (a single edge enters the ε0-regime in at most k0 steps). Let (ξ,η)∈Eα with ξ,η∈K. We claim that there is an integer n∈{0,…,k0} with d(Tn(ξ),Tn(η))≤ε0. Suppose not. Then d(Tn(ξ),Tn(η))>ε0 for every n∈{0,…,k0}, and since (Tn(ξ),Tn(η))∈Eα by Lemma 4.1, Lemma 4.2 yields
that is, k0≤D0ε0−1(1−φα(ε0))−1. This contradicts (13) and proves the claim.
Step 2 (the regime is absorbing). By Lemma 4.1 the sequence n↦d(Tn(ξ),Tn(η)) is non-increasing. Combined with Step 1 this shows that
d(Tn(ξ),Tn(η))≤ε0for every n≥k0 and every (ξ,η)∈Eα with ξ,η∈K.(15)
Step 3 (chaining). Let x,y∈K. Since N(α)<∞ there are an integer q≤N(α) and points ξ0,…,ξq∈K forming a Gˉα-chain from x to y. Fix n≥k0. By (C1) the points Tn(ξ0),…,Tn(ξq) again form a Gˉα-chain, and by (15) each of its steps satisfies
Since ε0=ηN(α)(ε)≤ηq(ε), Lemma 4.3 applies and gives d(Tn(x),Tn(y))≤ε. As x,y∈K were arbitrary, this proves (14) and hence (i).
Step 4 (the orbits are Cauchy). Let ρ>0 and choose ε>0 so small that U(ε)⊆{(u,v):d1(u,v)≤ρ}; replacing ε by min{ε,δ1(ε)} we may assume in addition that d(u,v)≤ε implies d(v,u)≤ε, so that (14) yields
(Tn(x),Tn(y))∈U(ε)(n≥k0(ε,α),x,y∈K).(16)
Given x∈K and m≥0, apply (16) with y:=Tm(x)∈K to get d1(Tn(x),Tn+m(x))≤ρ for all n≥k0(ε,α). Hence {Tn(x)}n≥1 is a Cauchy sequence in (X,d1). Since (X,d1) is complete and K is closed, the limit xx∗:=limn→∞Tn(x) exists and belongs to K.
Step 5 (the limit does not depend on the starting point). Let x,y∈K and ρ>0. By (16), d1(Tn(x),Tn(y))≤ρ for all sufficiently large n. Letting n→∞ and using the continuity of d1 we obtain d1(xx∗,xy∗)≤ρ, and since ρ>0 was arbitrary, xx∗=xy∗=:x∗. Finally, the integer k0(ε,α) of (13) does not depend on x, so the convergence is uniform on K. This proves (ii).
Corollary 5.3. Taking w≡1, φ1=φ and α=1 in Theorem 5.2 recovers [12 Theorem 2], with n0=N(1).
Theorem 5.2 asserts that all Picard orbits converge to a common limit, but it does not by itself identify that limit as a fixed point of T; the same is true of [12 Theorem 2]. The obstruction is that T is assumed to be contractive only along edges, so that no continuity of T at x∗ is available a priori. We record two hypotheses, either of which removes it. The second is the graph-theoretic one and is modelled on property (A) of [7].
Definition 5.4. The graded graph (G,w) has property(Aα) on K if, whenever {xn}n≥1⊆K satisfies (xn,xn+1)∈Eα for all n and xn→x in (X,d1) with x∈K, there exists a subsequence {xnk} with (xnk,x)∈Eα for all k.
Theorem 5.5 (Fixed point and uniqueness). Let Assumption 5.1 hold, let α∈A, and let x∗ be as in Theorem 5.2. Assume that at least one of the following conditions holds:
(a) T is continuous at x∗ as a mapping of (K,d1) into itself;
(b) (G,w) has property(Aα) on K and there exists x0∈K with (x0,T(x0))∈Eα.
ThenT(x∗)=x∗, andx∗ is the unique fixed point of T in K.
Proof. Throughout, x∗ denotes the point produced by Theorem 5.2(ii); we recall from that theorem that x∗∈K and that Tn(x)→x∗ in (X,d1) for every x∈K. The proof falls into three parts: the fixed point property under (a), the fixed point property under (b), and uniqueness, the last of which uses neither (a) nor (b).
Part I: T(x∗)=x∗ under hypothesis (a).
Step 1. Since x∗∈K, Theorem 5.2(ii) may be applied with the starting point x:=x∗, and gives
Tn(x∗)⟶x∗ in (X,d1).(17)
Step 2. The points Tn(x∗) all lie in K, because T maps K into itself. Hypothesis (a) states that T is continuous at x∗ as a mapping of (K,d1) into itself; applying this to the convergent sequence (17) gives
T(Tn(x∗))⟶T(x∗) in (X,d1).(18)
Step 3. On the other hand T(Tn(x∗))=Tn+1(x∗), and the sequence {Tn+1(x∗)}n≥0 is a shift of the sequence in (17); a shift of a convergent sequence converges to the same limit, so
T(Tn(x∗))⟶x∗ in (X,d1).(19)
Step 4.(X,d1) is a metric space, so a sequence has at most one limit. Comparing (18) and (19) gives T(x∗)=x∗.
Part II: T(x∗)=x∗ under hypothesis (b).
Let x0∈K be the point supplied by (b), so that (x0,T(x0))∈Eα, and put xn:=Tn(x0) for n≥0.
Step 1 (the orbit is a path in Eα). Apply Lemma 4.1 to the pair (x,y):=(x0,T(x0))∈Eα. Its first assertion gives (Tn(x0),Tn(T(x0)))∈Eα for every n≥0, that is,
(xn,xn+1)∈Eα(n≥0).(20)
(Equivalently one may argue directly by induction: if w(xn,xn+1)≥α then (C1) gives (xn+1,xn+2)∈E(G) together with $w(x_{n+1},x_{n+2})\ge w(x_n,x_{n+1})\ge\alpha.)
Step 2 (the orbit converges to x∗). By Theorem 5.2(ii),
xn⟶x∗ in (X,d1),x∗∈K.(21)
Step 3 (edges to the limit). Relations (20) and (21) are precisely the hypotheses of property (Aα) of Definition 5.4. That property therefore furnishes indices n1<n2<⋯ with
(xnk,x∗)∈Eα⊆E(G)(k≥1).(22)
This is the only place where hypothesis (b) is used, and it is what replaces the continuity assumed in Part I: without (22) the contractive estimate (8) could not be applied to any pair involving x∗.
Step 4 (from d1 to d). By (4), the d1-convergence (21) implies d(xn,x∗)→0. In particular, writing
ak:=d(xnk,x∗),bk:=d(xnk+1,x∗),
we have ak→0 and bk→0 as k→∞.
Step 5 (the images converge to T(x∗)). Fix k≥1. Both xnk and x∗ belong to K, and (xnk,x∗)∈E(G) by (22), so (8) applies to this pair and gives, using φ≤1 for every φ∈Φ,
d(T(xnk),T(x∗))≤φw(xnk,x∗)(ak)ak≤ak.
Since T(xnk)=xnk+1, this reads
ck:=d(xnk+1,T(x∗))≤ak⟶0.(23)
Step 6 (reversing the arguments). The distance d is not assumed symmetric, and (23) controls d(xnk+1,T(x∗)), whereas the triangle condition (A2) will be applied below to a chain whose first point is T(x∗). Let ε>0 and set
ε′:=δ2(ε)>0,ε′′:=δ1(ε′)>0.
By (23) there is k1 with ck≤ε′′ for all k≥k1. Assumption 2.1, applied with the target ε′ and the witness ε′′=δ1(ε′), then yields
d(T(x∗),xnk+1)≤ε′(k≥k1).(24)
Step 7 (applying the triangle condition). By Step 4 there is k2 with bk≤ε′ for all k≥k2. Fix any k≥max{k1,k2}. Then, by (24) and the definition of bk,
Assumption 2.2, applied to the triple (T(x∗),xnk+1,x∗), gives
d(T(x∗),x∗)≤ε.
Step 8. The number ε>0 in Steps 6 and 7 was arbitrary, so d(T(x∗),x∗)=0, and (1) gives T(x∗)=x∗.
Part III: uniqueness
Step 1. Let y∗∈K satisfy T(y∗)=y∗. An immediate induction gives Tn(y∗)=y∗ for every n≥0, so the sequence {Tn(y∗)}n≥0 is constant and converges in (X,d1) to y∗.
Step 2. Since y∗∈K, Theorem 5.2(ii) applied with the starting point y∗ gives Tn(y∗)→x∗ in (X,d1).
Step 3. Limits in the metric space (X,d1) are unique, so y∗=x∗. Combined with Parts I and II, this shows that x∗ is the unique fixed point of T in K. □
Remark 5.6. Part III of the preceding proof uses neither (a) nor (b), only Assumption 5.1 and Theorem 5.2. Thus a graded-edge Rakotch contraction satisfying Assumption 5.1 has at most one fixed point in K unconditionally; hypotheses (a) and (b) are needed only for existence. This is in contrast with [12 Theorem 1], where uniqueness requires the additional restriction that d be finite on K×K, a restriction subsumed here by (12).
Optimal threshold
The proof of Theorem 5.2 yields an explicit bound on the number of iterations needed to reach a prescribed accuracy. Since α ranges over A, the bound may be optimized over the threshold.
Theorem 6.1 (Iteration complexity). Let Assumption 5.1 hold and let ε>0. Then
x,y∈Ksupd(Tn(x),Tn(y))≤εfor every integer n≥K(ε),
Proof. Step 1 (each k0(ε,α) is a well-defined non-negative integer). Fix α∈A and abbreviate m:=N(α), which is finite by the definition of A. We check the three ingredients of (13) in turn.
(i) ηm(ε)>0. By (11), ηm(ε)=η1(δ2∘(m−1)(ε)). The moduli δ1,δ2 are witnesses supplied by Assumptions 2.1 and 2.2, hence take strictly positive values at strictly positive arguments. A finite induction of length m−1 therefore gives δ2∘(m−1)(ε)>0, and η1(t)=min{t,δ1(t)}>0 for t>0. It is here, and only here, that the finiteness of N(α) is used: for m=∞ the composition is an infinite one and the induction is unavailable.
(ii) 1−φα(ηm(ε))>0. Indeed ηm(ε)>0 by (i), and every member of Φ satisfies φ(t)<1 for t>0.
(iii) D0<∞ by Assumption 5.1.
Hence the quotient in (13) is a well-defined element of [0,∞), and k0(ε,α), being its ceiling, is a non-negative integer. (It equals 0 only in the degenerate case D0=0, in which K is a singleton.)
Step 2 (the infimum is attained). By Assumption 5.1 the set A is nonempty, so by Step 1
S:={k0(ε,α):α∈A}
is a nonempty subset of N∪{0}. By the well-ordering principle S has a least element, whence
K(ε)=α∈Ainfk0(ε,α)=minS
is a non-negative integer and there exists α∗∈A with
k0(ε,α∗)=K(ε).(25)
No monotonicity or continuity of α↦k0(ε,α) is claimed, and α∗ need not be unique; only the attainment of the least value is asserted, and that follows from the integrality of the k0 alone.
Step 3 (the conclusion of Theorem 5.2(i) does not involve α). This is the step that legitimises the optimization, so we isolate it. Write
Σn:=x,y∈Ksupd(Tn(x),Tn(y)).
Theorem 5.2(i) asserts: for every α∈A and every integer n≥k0(ε,α) one has Σn≤ε. The quantity Σn depends only on n, T and K; the threshold α enters the statement solely through the integer k0(ε,α) at which the conclusion switches on. Consequently the family of implications indexed by α∈A may be combined, and Σn≤ε holds as soon as n exceeds the smallest of these thresholds. Were the conclusion itself α-dependent, no such combination would be permitted and K(ε) would have to be replaced by a supremum.
Step 4 (conclusion). Let n≥K(ε) be an integer. By (25) we have n≥k0(ε,α∗), so Theorem 5.2(i), applied with the single threshold α:=α∗, gives Σn≤ε, which is the assertion. Finally, substituting ε0(ε,α)=ηN(α)(ε) from (13) into the definition of k0 yields the displayed formula for K(ε). □
We stress that Theorem 6.1 is not merely a restatement of Theorem 5.2(i) with a change of notation. The passage from a fixed α to the infimum over A is a genuine strengthening precisely because α↦k0(ε,α) is not monotone: by (13) its numerator is constant while its denominator involves ηN(α)(ε), which decreases as α grows, and 1−φα(⋅), which increases. Example 6.11 exhibits data for which the minimizing α∗ of (25) depends on ε, so that no single threshold is optimal for all tolerances.
For the remainder of this section we place ourselves in the quantitative regime of Remark 2.3 and we assume in addition that d is symmetric.
Remark 6.2. Symmetry is used only to ensure that the estimate (8), which is stated for a directed edge, may be applied to the reversed steps of a Gα-chain. Without it one obtains the same conclusion with an extra factor governed by a quantitative form of (A1); we do not pursue this here.
Lemma 6.3 (Polygonal inequality).Assume (5) and leta0,…,aq∈Xwithq≥1. Then
d(a0,aq)≤Mq−1i=0∑q−1d(ai,ai+1).
Proof. Induction on q. The case q=1 is trivial. If the assertion holds for q, then by (5) and M≥1,
Theorem 6.4 (Geometric rate with a graph constant).Let Assumption 5.1 hold, assume (5) and thatdis symmetric, and suppose that for someα∈Athe comparison function is constant, sayφα≡cα∈[0,1). Then, for every integern≥0,
x,y∈Ksupd(Tn(x),Tn(y))≤N(α)MN(α)−1D0cαn.
Proof. Let x,y∈K and let ξ0,…,ξq be a Gα-chain in K from x to y with q≤N(α). Fix i∈{0,…,q−1}. By (10) of Lemma 4.1,
d(Tn(ξi),Tn(ξi+1))≤cαnd(ξi,ξi+1)≤cαnD0,
where symmetry of d has been used to cover reversed steps. Applying Lemma 6.3 to the points Tn(ξ0),…,Tn(ξq) we obtain
Now we exhibit three graded graphs on which every quantity occurring in § 5 and § 7, the grade of each edge, the level cuts Eα, the chain diameter N(α), the admissible set A, the Picard iterates, and the estimates of Theorems 6.1 and 6.4, is computed explicitly. Everything is phrased in the language of the weighted graph: vertices, edges carrying a grade w(e)∈(0,1], level cuts, chains, and the graded Rakotch condition (8).
Throughout, X=R and
d(x,y)=∣x−y∣p(p≥1),(26)
so that, by the convexity of t↦tp, condition (5) holds with M=2p−1; the induced uniformity is the Euclidean one, d1 may be taken to be the Euclidean metric, and (X,d1) is complete. In each example K⊂R is closed and bounded, so Assumption 5.1 is met once A=∅ is verified. All gradings are the canonical ones of Proposition 3.4, so that
w(x,y)=1−c(x,y),φα≡cα=1−α,(27)
with c(x,y) the supremum of the orbital ratios (9).
We first record an elementary structural fact which explains the shape of the tables below: on a finite graded graph the Picard iteration does not merely converge, it terminates.
Proposition 6.5 (Finite graded graphs collapse). Let Assumption 5.1 hold and suppose K is finite with at least two points. Then there is an integer n†≥1 such that Tn is constant on K for every n≥n†; that is, Tn(x)=x∗ for all x∈K and all n≥n†.
Proof. Put δ:=min{d(x,y):x,y∈K,x=y}, which is positive because K is finite and d satisfies (1). Choose ε∈(0,δ) and let n†:=k0(ε,α) for some α∈A, as in (13). By Theorem 5.2(i), d(Tn(x),Tn(y))≤ε<δ for all x,y∈K and all n≥n†. By the definition of δ this forces Tn(x)=Tn(y), so Tn is constant; its value is the common limit x∗ of Theorem 5.2(ii).
Consequently, on a finite graded graph the geometric estimate of Theorem 6.4 is an envelope which is eventually strictly larger than the (vanishing) true quantity. What one should compare is therefore not the two sequences but the two iteration counts: the number of steps predicted by Theorem 6.1 against the number actually required. Example 6.6 shows that the graded prediction can be exact.
Example 6.6. Let p=2, so that M=2, and let
K={v0,v1,v2,v3,v4},v0=0,vi=3i−4(1≤i≤4),
that is K={0,271,91,31,1}, with
T(v0)=v0,T(vi)=vi−1(1≤i≤4).
Let G be the complete graph on K (all 10 unordered pairs, together with all loops), so that the edge-preservation half of (C1) is automatic.
Since {vi} is decreasing and T shifts indices downwards, the orbital ratios are computed at once; writing Δ=j−i, one finds for 1≤i<j≤4
The resulting canonical grades (27) are collected in Table 1 for Example 6.6; they take five distinct values. Grade monotonicity (C1) holds on every edge, as guaranteed by Proposition 3.4.
edge class
pairs
c(⋅,⋅)
grade w
{v0,v1}
v0v1
0
1
spokes at v0
v0v2,v0v3,v0v4
1/9
8/9≈0.8889
Δ=3
v1v4
81/676
595/676≈0.8802
Δ=2
v1v3,v2v4
9/64
55/64≈0.8594
Δ=1
v1v2,v2v3,v3v4
1/4
3/4=0.7500
Table 1. Canonical Grades
Figure 1. Figure 1, shows the graded graph of Example 6.6. Thick edges carry grade ≥8/9, dashed edges grades 55/64 and 595/676, dotted edges grade 3/4. The mapping T slides each vertex one step towards v0.
Level cuts and the admissible set. Ordering the grades as in Table 1 gives
Indeed, for α≤43 every pair is an edge of Gα, so N(α)=1; for 43<α≤98 the three Δ=1 edges are deleted, but v0 remains adjacent to every vertex, so any two vertices are joined by a Gα-chain of two hops through v0 and N(α)=2; for α>98 only {v0,v1} survives and Gα is disconnected. Figure 2 displays the two admissible cuts of Example 6.6. Raising the threshold from 43 to 98 improves the contraction factor from cα=41 to cα=91 but doubles the chain diameter.
Figure 2. Gα for α=43: N(α)=1; Gα for α=98: N(α)=2.
Iterates. The orbit of the vertex set is
n=0:{0,271,91,31,1},n=1:{0,0,271,91,1},
n=2:{0,0,0,271,91},n=3:{0,0,0,0,271},
n=4:{0}, so T4 is constant, in accordance with Proposition 6.5, and the unique fixed point is x∗=v0=0 (Theorem 5.5(a), T being continuous). With D0=supx,y∈Kd(x,y)=1 one obtains
x,y∈Ksupd(Tn(x),Tn(y))={9−n0(0≤n≤3),(n≥4).
Rates and the optimal threshold.Theorem 6.4 gives, at the two admissible thresholds,
and Λ(α2,ε)<Λ(α1,ε) precisely when log(1/ε)>2(log2)2/log(3/2), that is, when ε<ε∗ with
ε∗=exp(−log(3/2)2(log2)2)≈0.0935.
Thus the optimal threshold is α1=43 for coarse tolerances and α2=98 for fine ones; since α1 is exactly the ungraded threshold α0=inf{w(e)} of Remark 3.5, the graded theory is strictly better than the ungraded one for every ε<ε∗. Table 2 makes this concrete. The least n for which the bound of Theorem 6.4 certifies supx,yd(Tnx,Tny)≤ε, against the least n for which this actually holds, in Example 6.6. At the graded threshold α2 the prediction is exact; the ungraded threshold α1 overestimates by one step.
ε
steps predicted at α1=43
steps predicted at α2=98
steps actually needed
10−1
2
2
2
10−2
4
3
3
10−3
5
4
4
Table 2. Predicted vs Actual Steps
Remark 6.7. The agreement in the last two columns of Table 2 is not an accident of rounding: at α2 one has 4⋅9−3=0.005487≤10−2 while 4⋅9−2=0.04938>10−2, and the true values are 9−3=0.001372 and 9−2=0.01235. The graded estimate therefore locates the exact stopping time, whereas the ungraded estimate 4−n requires n=4.
The previous example has N(α)≤2, so the factor MN(α)−1 of Theorem 6.4 never exceeds M. We now exhibit a graded graph which is not complete and whose admissible cuts have N(α)=3, so that the graph constant N(α)MN(α)−1 genuinely reflects the length of the chains used in Step 3 of the proof of Theorem 5.2.
Example 6.8. Let p=2, so M=2, and let
K={x0,x1,x2,x3}={0,1,3,9},
T(x0)=x0,T(xi)=xi−1(1≤i≤3),
with E(G) consisting of the loops together with the three consecutive pairs {x0,x1}, {x1,x2}, {x2,x3} and their reversals. The edge set is T-invariant, since {xi,xi+1}↦{xi−1,xi} for i≥1 and {x0,x1} maps to a loop. The orbital ratios are
c(x0,x1)=0,c(x1,x2)=c(x2,x3)=41,
whence the grades w(x0,x1)=1 and w(x1,x2)=w(x2,x3)=43. Therefore
N(α)=3for 0<α≤43,N(α)=∞for α>43,A=(0,43],
the cut at α=43 being the whole path x0−x1−x2−x3, of graph diameter 3, and the cut above 43 retaining only {x0,x1}, which leaves x2,x3 isolated. Figure 3 presents the graded path of Example 6.8; the grey arrows indicate the action of T. The admissible cut has chain diameter N(α)=3.
Figure 3.
Here D0=∣9−0∣2=81 and, at the optimal threshold α=43 with cα=41, Theorem 6.4 yields
against the true values 81,9,1,0,0,… for n=0,1,2,3,…. The estimate is valid at every step but is conservative, by a factor N(α)MN(α)−1=12 coming from the three-hop chains and the constant M=2 of the relaxed triangle inequality. This is the price, quantified by Lemma 6.3, of comparing d with the sum of the d-lengths of the hops of a chain; Example 6.9 shows that when N(α)=1 no such price is paid and the estimate becomes an identity.
Example 6.9. Let p=1, so that d(x,y)=∣x−y∣ and M=1, and let
K={0}∪{2−n:n≥0},T(x)=2x,
with G the complete graph on K.
The set K is closed and bounded in R, with D0=1, and T(K)⊆K. Because T is linear, every orbital ratio equals 21: for all x,y∈K,
d(x,y)d(T(x),T(y))=∣x−y∣∣x/2−y/2∣=21,
so the canonical grading is constant, w=21, and cα=21 for every α∈A=(0,21]. Since G is complete, N(α)=1 throughout A. Figure 4 exhibits the infinite graded ray of Example 6.9 (complete graph; only some edges drawn). Every edge carries grade w=21, and N(α)=1.
Theorem 6.4 therefore predicts
x,y∈Ksupd(Tn(x),Tn(y))≤N(α)MN(α)−1D0cαn=2−n,
while directly supx,y∈K∣Tn(x)−Tn(y)∣=2−nsupx,y∈K∣x−y∣=2−n. Hence
x,y∈Ksupd(Tn(x),Tn(y))=N(α)MN(α)−1D0cαn(n≥0):
the estimate of Theorem 6.4 is attained with equality for every n. In particular the constant N(α)MN(α)−1D0 cannot be improved in general. Note also that K is infinite, so Proposition 6.5 does not apply and no orbit terminates: the geometric decay persists for all n, and Tn(x)→0=x∗ without ever reaching it for $x\ne0.
Figure 4.
Corollary 6.10 (Optimal threshold). Under the hypotheses ofTheorem 6.4, suppose thatφα≡cα∈[0,1)for everyα∈A. Then, for everyε∈(0,D0),
Proof. Solve N(α)MN(α)−1D0cαn≤ε for n and use Theorem 6.4.
The numerator of Λ(⋅,ε) is non-decreasing in α, because N(α) is; the denominator is likewise non-decreasing, because cα is non-increasing. The infimum is therefore attained, in general, at neither endpoint of A. The following example makes this precise, and shows in addition that the ungraded choice α0 of Remark 3.5 is strictly suboptimal.
Example 6.11 (An interior optimum). Fix p>1 and b∈(0,1). Let X=R with
d(x,y):=∣x−y∣p,
let K:=[0,b] and let
T(x):=21x2(x∈K).
Let G be the complete graph on X, so that E(G)=X×X.
We verify the hypotheses and compute all the quantities involved.
(1) The space. By the convexity of t↦tp on [0,∞),
∣x−z∣p≤(∣x−y∣+∣y−z∣)p≤2p−1(∣x−y∣p+∣y−z∣p),
so (5) holds with M=2p−1>1, and d is symmetric and satisfies (1). Since U(ε)={(x,y):∣x−y∣≤ε1/p}, the uniformity (3) is the usual one on R, so d1 may be taken to be the Euclidean metric and (X,d1) is complete. The set K is closed and D0=bp.
(2) The mapping. If 0≤x≤b<1 then 0≤x2/2≤b/2≤b, so T(K)⊆K. For x,y∈K,
Note that α1=inf{w(e):e∈E(G),e not a loop} is exactly the ungraded threshold α0 of Remark 3.5, and that 0<α1<α2<1.
(5) The optimization. Since Λ(⋅,ε) of (28) is minimized, on each of the two branches, at the largest admissible α, it suffices to compare α1 and α2. Write B:=log(D0/ε)=log(bp/ε)>0. Using M=2p−1 and logM+log2=plog2,
Λ(α1,ε)=plog(1/b)B,Λ(α2,ε)=plog(2/b)B+plog2.
Setting β:=log(1/b)>0, the inequality Λ(α2,ε)<Λ(α1,ε) is equivalent to B(log2+β)>β(B+plog2), that is, to B>pβ. Hence
α∈AargminΛ(α,ε)={α1,α2,ε>b2p,ε<b2p,
with equality of the two values at ε=b2p.
Remark 6.12. Example 6.11 settles the objection of Remark 3.5 in both of its aspects.
(i) The optimum is interior. For ε∈(b2p,D0) the minimizer is α1, which lies strictly inside A=(0,α2]. The extreme choices are both bad: as α↓0 one has cα↑1 and Λ(α,ε)→∞, so the weakest threshold is useless, while for ε>b2p the strongest admissible threshold α2 is strictly worse than α1.
(ii) The grading is not cosmetic. For ε<b2p the graded optimum α2 is strictly better than the ungraded choice α0=α1, and
ε↓0limΛ(α1,ε)Λ(α2,ε)=log(2/b)log(1/b)<1.
For b=21 this ratio equals 21: the graded bound is asymptotically half the ungraded one. Since the two bounds are derived from the same theorem applied at two thresholds, the improvement is a strict gain of the graded formulation over the formulation of [12], not an artefact of a different proof.
The example is a snowflake b-metric space: with p>1 it carries the genuine constant M=2p−1>1, so the factor MN(α)−1 of Theorem 6.4 is active. The interior optimum is nevertheless not an artefact of this structure. Repeating the computation in the underlying metric D(x,y)=∣x−y∣, that is with p=1 and M=1, leaves the combinatorics of the level cuts unchanged, replaces the crossover value b2p by b2, and leaves the asymptotic ratio log(1/b)/log(2/b) unchanged; the trade-off is therefore driven by the chain-length factor N(α) alone and survives in an ordinary metric space. Only two values of N(α) occur here, so the trade-off is resolved by comparing two branches rather than by differentiating a smooth function. Constructions with N(α) taking any prescribed finite set of values may be obtained by replacing K with a finite union of suitably scaled copies of [0,b] and deleting the edges between distinct copies; whether N(α) can be made to grow along a continuum, so that the optimum is characterized by a first-order condition.
Applications
A quadratic integral equation
Let k:[0,1]2→[0,∞) be continuous with ∫01k(t,s)ds≤1 for every t, let b∈(0,1) and let p>1. Consider the equation
x(t)=21∫01k(t,s)x(s)2ds,t∈[0,1].(30)
Proposition 7.1. Equation (30) has a unique solution x∗ in K:={x∈C[0,1]:0≤x≤b}, namely x∗≡0, and for every x∈K the Picard iterates of the right-hand side converge to x∗ uniformly on [0,1], with
Proof. Let X=C[0,1] with d(x,y)=∥x−y∥∞p; as in Example 6.11, (5) holds with M=2p−1, the induced uniformity is that of uniform convergence, (X,d1) is complete and K is closed with D0=bp. If x∈K then 0≤(Fx)(t)≤21∥x∥∞2≤b/2≤b, so F(K)⊆K. For x,y∈K and t∈[0,1],
whence d(Fx,Fy)≤(∥x+y∥∞/2)pd(x,y). Take G complete and w(x,y):=1−(∥x+y∥∞/2)p. Condition (C1) holds because, using ∫01k(t,s)ds≤1 and x2+y2≤b(x+y) pointwise on [0,b],
so that ∥Fx+Fy∥∞≤∥x+y∥∞ and hence w(Fx,Fy)≥w(x,y). Thus F is a graded-edge Rakotch contraction with φα≡1−α. Choosing α=α2=1−(b/2)p, so that N(α)=2 and cα=(b/2)p, Theorem 6.4 gives supx,y∈Kd(Fnx,Fny)≤2⋅2p−1⋅bp⋅(b/2)pn, and Theorem 5.5(a) applies since F is continuous. The constant function 0 is a fixed point, hence is the unique one. □
Remark 7.2. The solution of (30) in K being trivial, the interest of Proposition 7.1 lies in the rate rather than in the existence, and in fact that the rate is obtained from a graded threshold which is not the ungraded one. A non-trivial variant is obtained by adding an inhomogeneous term g∈C[0,1] with 0≤g≤b/2 to the right-hand side of (30); the same computation applies verbatim provided 21b2+∥g∥∞≤b, and the fixed point is then a non-trivial solution.
A finite weighted digraph
For a finite K the invariant N(α) is computable directly: it is the diameter of the level-cut graph Gα, obtainable by breadth-first search in O(∣K∣∣Eα∣) time, and the admissible set A is determined by the ∣E(G)∣ distinct values taken by w. The optimization in Corollary 6.10 is then a finite comparison over at most ∣E(G)∣ candidate thresholds, and α∗ may be tabulated exactly. We omit the numerical illustration.
Conclusion
The framework of this paper separates two functions that the graph performs in [12]: connecting the space, and carrying the contraction data. Once separated, they are seen to be in competition, and the resulting optimization is what the grading buys. In future we plan to construct a generalized metric space, a graded graph and a mapping satisfying Assumption 5.1 for which α↦N(α) is unbounded on A and takes infinitely many values, so that the minimization in Corollary 6.10 is governed by a first-order condition rather than by a comparison of finitely many branches. Example 6.11 realizes two branches; the difficulty in realizing a continuum is that the contraction ratio c(x,y) of Proposition 3.4 depends on the position of the pair in the space, whereas N(α) depends on the combinatorial length of chains, and coupling the two requires a graph whose edge lengths and edge grades are simultaneously prescribed.
Author Contributions All authors contributed equally and significantly in writing this paper. All authors read and approved the final manuscript.
Funding Not applicable.
Data Availability Not applicable.
Declarations
Competing interests The authors declare that they have no competing interests.
References
[1]I. A. Bakhtin, The contraction mapping principle in quasimetric spaces, Funct. Anal., UI’yanovsk. Gos. Ped. Inst. 30 (1989), 26–37.
[2]S. Banach, Sur les opérations dans les ensembles abstraits et leur application aux équations intégrales, Fund. Math. 3 (1922), 133–181.DOI
[3]I. Beg, A.R. Butt and S. Radojevic, Contraction principle for set valued mappings on a metric space with a graph, Computers and Math. with Applications, 60(2010), 1214-1219DOI
[4]I. Beg and A.R. Butt: Fixed points of set valued graph contractions, J. Inequalities and Appl., 2013(2013):252, 7 pages
[5]F. Bojor, Fixed point theorems for Reich type contractions on metric spaces with a graph, Nonlinear Anal. 75 (2012), 3895–3901.DOI
[7]J. Jachymski, The contraction principle for mappings on a metric space with a graph, Proc. Amer. Math. Soc. 136 (2008), 1359–1373.DOI
[8]J. J. Nieto and R. Rodríguez-López, Contractive mapping theorems in partially ordered sets and applications to ordinary differential equations, Order 22 (2005), 223–239.DOI
[9]E. Rakotch, A note on contractive mappings, Proc. Amer. Math. Soc. 13 (1962), 459–465.DOI
[10]A. C. M. Ran and M. C. B. Reurings, A fixed point theorem in partially ordered sets and some applications to matrix equations, Proc. Amer. Math. Soc. 132 (2004), 1435–1443.DOI
[11]S. Reich and A. J. Zaslavski, A fixed point result in generalized metric spaces, J. Anal. 30 (2022), 1467–1473.DOI
[12]S. Reich and A. J. Zaslavski, A fixed point result in generalized metric spaces with graphs, Ann. Acad. Rom. Sci. Ser. Math. Appl. 18 (2026), no. 1, 9–18.DOI