Logically Contractive Mappings: Fixed Points and Event-Indexed Rates
Faruk Alpay, Taylan Alpay · 2025
Read Logically Contractive Mappings: Fixed Points and Event-Indexed Rates by Faruk Alpay, Taylan Alpay (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-07059.txt.
Introduction
A classic math result says a map that always shrinks distances must settle at one fixed point. Faruk Alpay and Taylan Alpay loosen that rule in this 2025 paper. Their maps may hold distance steady most of the time, as long as they shrink it along some steps in the chain. They call such maps logically contractive.
The authors prove a fixed point result that widens the classic Banach principle. They give speed rates tied to the steps where shrinking happens, called events. When gaps between events stay bounded, the rates become plain step counts. One worked case shows a map whose square shrinks strictly even though the map itself does not. The results tie together several older ideas about loose shrinking maps and raise new questions about sparse events.
Catalog record
| CREATOR | Faruk Alpay, Taylan Alpay |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 17 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-07059
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Logically Contractive Mappings: Fixed Points and Event-Indexed Rates Faruk Alpay, Taylan Alpay Abstract We introduce logically contractive mappings —nonexpansive self-maps that contract along a subsequence of iterates—and prove a fixed-point theorem that extends Banach’s principle. We obtain event-indexed convergence rates and, under bounded gaps between events, explicit iteration-count rates. A worked example shows a nonexpansive map whose square is a strict contraction, and we clarify relations to Meir–Keeler and asymptotically nonexpansive mappings. We further generalize to variable-factor events and show that ∏ k λ k = 0 \prod_{k}\lambda_{k}=0 (equivalently ∑ k − ln λ k = ∞ \sum_{k}-\ln\lambda_{k}=\infty ) implies convergence. These results unify several generalized contraction phenomena and suggest new rate questions tied to event sparsity. 1 Introduction Banach’s contraction mapping theorem is a cornerstone of metric fixed-point theory: any map T on a complete metric space (X,d) with a uniform Lipschitz constant L<1 has a unique fixed point z\in X , and moreover T^{n}x converges to z for every x\in X with an exponential rate L^{n}d(x,z) . Numerous extensions have been studied, weakening the strict contraction requirement while retaining fixed-point existence and convergence guarantees. Classic examples include Meir–Keeler contractions [ 2 ] (which relax uniform L<1 to a certain \varepsilon - \delta decay condition) and asymptotically nonexpansive maps (which allow the Lipschitz constant of T^{n} to approach 1 as n\to\infty ) [ 3 ] . In this note, we formalize a simple unifying concept: a mapping that is overall nonexpansive ( \Lip(T)\leq 1 ) but occasionally acts like a contraction. Specifically, a logically contractive mapping is one for which there exists some 0<\lambda<1 and an infinite sequence of iterate counts n_{1}<n_{2}<n_{3}<\cdots such that each T^{n_{k}} is \lambda^{k} -Lipschitz. Intuitively, T is not required to contract every time step, but it does so at an infinite sequence of “event” iterations, eventually forcing convergence. We prove that any logically contractive map has a unique fixed point and that all iterations converge to it (Theorem 2.2 ). In fact, the distance to the fixed point decays at least geometrically along the event subsequence, and if the gap between consecutive events is bounded, one can derive an explicit per-iteration decay rate (Corollary 2.4 ). We also compare logically contractive maps with the Meir–Keeler and asymptotically nonexpansive classes, showing that none of these notions subsume the others (Proposition 3.1 ). A concrete example is provided (Example 3.2 ) of a nonexpansive mapping that is not a strict contraction yet is logically contractive (in fact, its second iterate is constant), illustrating the concept. Finally, we generalize the logical contractiveness idea to allow varying contraction factors at different events. We show that if the product of these factors tends to zero (equivalently, an infinite sum of their logarithmic decreases diverges), then T^{n}x still converges to the fixed point for any starting point x (Theorem 4.2 ). We conclude with brief remarks on the scope of our results and open questions. Notation. Throughout, (X,d) is a complete metric space. For a map T:X\to X and any n\in\mathbb{N} , T^{n} denotes the n -fold composition (so T^{1}=T ). We write \Lip(S) for the (global) Lipschitz constant of a function S , i.e. \Lip(S):=\sup_{x\neq y}\frac{d(Sx,Sy)}{d(x,y)} . A map is nonexpansive if \Lip(T)\leq 1 . Compositions of nonexpansive maps are nonexpansive as well, i.e. \Lip(T^{n})\leq 1 for any n if \Lip(T)\leq 1 . We write \lfloor t\rfloor for the floor of t (the greatest integer \leq t ). 2 Fixed-point theorem for logically contractive mappings Definition 2.1 (Logically contractive mapping) . Let T:X\to X be a self-map on a metric space (X,d) . We say T is logically contractive if T is nonexpansive and there exist a constant \lambda\in(0,1) and an increasing sequence of natural numbers n_{1}<n_{2}<n_{3}<\cdots such that for all k\geq 1 , \Lip\!\big(T^{\,n_{k}}\big)\;\leq\;\lambda^{k}. In other words, T^{\,n_{k}} is a strict contraction, with Lipschitz constant at most \lambda^{k} , at an infinite sequence of iterate indices n_{k} . By definition, a logically contractive map T has at least one “event” (the case k=1 ) where a strict contraction occurs. In particular, \Lip(T^{\,n_{1}})\leq\lambda<1 , so T^{\,n_{1}} is itself a Banach contraction on X . Therefore T^{\,n_{1}} has a unique fixed point in X , which we will denote z : T^{\,n_{1}}(z)=z. Theorem 2.2 . Let T:X\to X be a logically contractive mapping. Then: 1. T has a unique fixed point z\in X . 2. For every x\in X , the sequence T^{n}x converges to z as n\to\infty . 3. Moreover, for every x\in X and all k\geq 1 , d\!\big(T^{\,n_{k}}x,\;z\big)\;\leq\;\lambda^{k}\,d(x,z). Proof. (1) Existence/uniqueness. Since \Lip(T^{\,n_{1}})\leq\lambda<1 , T^{\,n_{1}} is a Banach contraction and hence has a unique fixed point z\in X . Now T^{\,n_{1}}(Tz)=T(T^{\,n_{1}}z)=Tz , so by uniqueness of the fixed point of T^{\,n_{1}} we have Tz=z . If Tx=x for some x , then T^{\,n_{1}}x=x , whence x=z by uniqueness. (2) Convergence of the full sequence. By (3) below, d(T^{\,n_{k}}x,z)\leq\lambda^{k}d(x,z)\to 0 as k\to\infty . If n_{k}\leq n<n_{k+1} , then by nonexpansiveness of T^{\,n-n_{k}} , d(T^{n}x,z)\;=\;d\!\big(T^{\,n-n_{k}}(T^{\,n_{k}}x),\;T^{\,n-n_{k}}(z)\big)\;\leq\;d(T^{\,n_{k}}x,z). Hence d(T^{n}x,z)\to 0 as n\to\infty . (3) Eventwise estimate. Since T^{\,n_{k}}z=z and \Lip(T^{\,n_{k}})\leq\lambda^{k} , d(T^{\,n_{k}}x,z)\;=\;d\!\big(T^{\,n_{k}}x,\;T^{\,n_{k}}z\big)\;\leq\;\lambda^{k}\,d(x,z). ∎ Remark 2.3 (Role of completeness) . Completeness of (X,d) is used only to invoke Banach’s theorem for the strict contraction T^{\,n_{1}} ; all subsequent steps rely on nonexpansiveness and the eventwise Lipschitz bounds. From part (3) we obtain a convergence rate in terms of the event count k . To express a per-iteration rate, we need information on the spacing of events. Corollary 2.4 (Iteration-count rate under bounded event gaps) . Assume T is logically contractive with constant \lambda\in(0,1) and let z be the fixed point of T . If there is a uniform bound M on event gaps (i.e. n_{k+1}-n_{k}\leq M for all k ), then for every x\in X and every n\geq n_{1} , d\!\big(T^{n}x,\;z\big)\;\leq\;\lambda^{\,1+\left\lfloor\frac{\,n-n_{1}\,}{M}\right\rfloor}\,d(x,z). Proof. Let m:=\left\lfloor\frac{n-n_{1}}{M}\right\rfloor . Then n_{1}+mM\leq n<n_{1}+(m+1)M . By induction from n_{k+1}\leq n_{k}+M we have n_{m+1}\leq n_{1}+mM , hence n_{m+1}\leq n . By Theorem 2.2 (3) and nonexpansiveness, d(T^{n}x,z)\;\leq\;d(T^{\,n_{m+1}}x,z)\;\leq\;\lambda^{\,m+1}d(x,z). ∎ We now show that a bounded-gap schedule can always be enforced by using the first strict event repeatedly. Lemma 2.5 (Bounded gaps without loss of generality) . Let T be logically contractive with constant \lambda and first event time n_{1} . Then for every integer m\geq 1 , \Lip\!\big(T^{\,mn_{1}}\big)\;=\;\Lip\!\big((T^{\,n_{1}})^{m}\big)\;\leq\;\big(\Lip(T^{\,n_{1}})\big)^{m}\;\leq\;\lambda^{m}. Thus the canonical schedule \tilde{n}_{m}:=mn_{1} certifies logical contractiveness with constant gap \tilde{n}_{m+1}-\tilde{n}_{m}=n_{1} . Proof. The identity (T^{\,n_{1}})^{m}=T^{\,mn_{1}} is equality of maps. Submultiplicativity under composition gives \Lip\!\big((T^{\,n_{1}})^{m}\big)\;\leq\;\big(\Lip(T^{\,n_{1}})\big)^{m}, and equality need not hold in general (indeed, strict inequality is typical). ∎ Corollary 2.6 (Explicit per-iterate rate) . For any logically contractive T with first event time n_{1} and fixed point z , we have for all x\in X and all n\geq 0 , d\!\big(T^{n}x,\;z\big)\;\leq\;\big(\Lip(T^{\,n_{1}})\big)^{\left\lfloor\frac{n}{\,n_{1}\,}\right\rfloor}\,d(x,z)\;\leq\;\lambda^{\left\lfloor\frac{n}{\,n_{1}\,}\right\rfloor}\,d(x,z). Proof. Apply Corollary 2.4 to the canonical schedule \tilde{n}_{m}=mn_{1} with gap M=n_{1} . ∎ Remark 2.7 . The original schedule \{n_{k}\} may be irregular; the canonical schedule in Lemma 2.5 produces a clean rate. If later events have much smaller Lipschitz constants than \lambda , one can restart with a smaller contraction constant to sharpen the bound. Remark 2.8 (Equivalent formulation) . A mapping T is logically contractive if and only if it is nonexpansive and there exists at least one iterate T^{N} that is a strict contraction (i.e. \Lip(T^{N})<1