Results 1  10
of
11
Local stability of ergodic averages
 Transactions of the American Mathematical Society
"... We consider the extent to which one can compute bounds on the rate of convergence of a sequence of ergodic averages. It is not difficult to construct an example of a computable Lebesguemeasure preserving transformation of [0, 1] and a characteristic function f = χA such that the ergodic averages An ..."
Abstract

Cited by 27 (4 self)
 Add to MetaCart
We consider the extent to which one can compute bounds on the rate of convergence of a sequence of ergodic averages. It is not difficult to construct an example of a computable Lebesguemeasure preserving transformation of [0, 1] and a characteristic function f = χA such that the ergodic averages Anf do not converge to a computable element of L2([0,1]). In particular, there is no computable bound on the rate of convergence for that sequence. On the other hand, we show that, for any nonexpansive linear operator T on a separable Hilbert space, and any element f, it is possible to compute a bound on the rate of convergence of (Anf) from T, f, and the norm ‖f ∗ ‖ of the limit. In particular, if T is the Koopman operator arising from a computable ergodic measure preserving transformation of a probability space X and f is any computable element of L2(X), then there is a computable bound on the rate of convergence of the sequence (Anf). The mean ergodic theorem is equivalent to the assertion that for every function K(n) and every ε> 0, there is an n with the property that the ergodic averages Amf are stable to within ε on the interval [n, K(n)]. Even in situations where the sequence (Anf) does not have a computable limit, one can give explicit bounds on such n in terms of K and ‖f‖/ε. This tells us how far one has to search to find an n so that the ergodic averages are “locally stable ” on a large interval. We use these bounds to obtain a similarly explicit version of the pointwise ergodic theorem, and show that our bounds are qualitatively different from ones that can be obtained using upcrossing inequalities due to Bishop and Ivanov. Finally, we explain how our positive results can be viewed as an application of a body of general prooftheoretic methods falling under the heading of “proof mining.” 1
On the computational content of the Krasnoselski and Ishikawa fixed point theorems
, 2000
"... This paper is a case study in proof mining applied to noneffective proofs in nonlinear functional analysis. More specifically, we are concerned with the fixed point theory of nonexpansive selfmappings f of convex sets C in normed spaces. We study the Krasnoselski iteration as well as more general ..."
Abstract

Cited by 11 (10 self)
 Add to MetaCart
This paper is a case study in proof mining applied to noneffective proofs in nonlinear functional analysis. More specifically, we are concerned with the fixed point theory of nonexpansive selfmappings f of convex sets C in normed spaces. We study the Krasnoselski iteration as well as more general socalled KrasnoselskiMann iterations. These iterations converge to fixed points of f under certain compactness conditions. But, as we show, already for uniformly convex spaces in general no bound on the rate of convergence can be computed uniformly in f . This is related to the nonuniqueness of fixed points. However, the iterations yield even without any compactness assumption and for arbitrary normed spaces approximate fixed points of arbitrary quality for bounded C (asymptotic regularity, Ishikawa 1976). We apply proof theoretic techniques (developed in previous papers of us) to noneffective proofs of this regularity and extract effective uniform bounds on the rate of the asymptotic re...
The Use of a Logical Principle of Uniform Boundedness in Analysis
, 1996
"... This paper is part of a sequence of papers ([9],[10],[11],[12]) resulting from our Habilitation thesis [8] addressing the following question: What is the impact on the growth of extractable uniform bounds the use of various analytical principles \Gamma in a given proof of an 89sentence might have? ..."
Abstract

Cited by 9 (8 self)
 Add to MetaCart
This paper is part of a sequence of papers ([9],[10],[11],[12]) resulting from our Habilitation thesis [8] addressing the following question: What is the impact on the growth of extractable uniform bounds the use of various analytical principles \Gamma in a given proof of an 89sentence might have? In particular we are interested in analyzing proofs of sentences having the form (1) 8u
On the Arithmetical Content of Restricted Forms of Comprehension, Choice and General Uniform Boundedness
 PURE AND APPLIED LOGIC
, 1997
"... In this paper the numerical strength of fragments of arithmetical comprehension, choice and general uniform boundedness is studied systematically. These principles are investigated relative to base systems T n in all finite types which are suited to formalize substantial parts of analysis but ..."
Abstract

Cited by 9 (4 self)
 Add to MetaCart
In this paper the numerical strength of fragments of arithmetical comprehension, choice and general uniform boundedness is studied systematically. These principles are investigated relative to base systems T n in all finite types which are suited to formalize substantial parts of analysis but nevertheless have provably recursive function(al)s of low growth. We reduce the use of instances of these principles in T n proofs of a large class of formulas to the use of instances of certain arithmetical principles thereby determining faithfully the arithmetical content of the former. This is achieved using the method of elimination of Skolem functions for monotone formulas which was introduced by the author in a previous paper. As
On the Computational Content of the BolzanoWeierstraß Principle
, 2009
"... We will apply the methods developed in the field of ‘proof mining’ to the BolzanoWeierstraß theorem BW and calibrate the computational contribution of using this theorem in proofs of combinatorial statements. We provide an explicit solution of the Gödel functional interpretation (combined with nega ..."
Abstract

Cited by 4 (4 self)
 Add to MetaCart
We will apply the methods developed in the field of ‘proof mining’ to the BolzanoWeierstraß theorem BW and calibrate the computational contribution of using this theorem in proofs of combinatorial statements. We provide an explicit solution of the Gödel functional interpretation (combined with negative translation) as well as the monotone functional interpretation of BW for the product space ∏i∈N[−k i, k i] (with the standard product metric). This results in optimal program and bound extraction theorems for proofs based on fixed instances of BW, i.e. for BW applied to fixed sequences in ∏i∈N[−k i, k i].
Things that can and things that can't be done in PRA
, 1998
"... It is wellknown by now that large parts of (nonconstructive) mathematical reasoning can be carried out in systems T which are conservative over primitive recursive arithmetic PRA (and even much weaker systems). On the other hand there are principles S of elementary analysis (like the BolzanoW ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
It is wellknown by now that large parts of (nonconstructive) mathematical reasoning can be carried out in systems T which are conservative over primitive recursive arithmetic PRA (and even much weaker systems). On the other hand there are principles S of elementary analysis (like the BolzanoWeierstra principle, the existence of a limit superior for bounded sequences etc.) which are known to be equivalent to arithmetical comprehension (relative to T ) and therefore go far beyond the strength of PRA (when added to T ). In this paper
Proof Mining in Practice
, 2008
"... In this paper, we present some aspects of a recent application of proof mining by J.Avigad, H.Towsner and the author. In this case study, we analysed a proof of the Mean Ergodic Theorem and obtained a computable rate of convergence for the ergodic averages. Proof mining generally falls into two main ..."
Abstract
 Add to MetaCart
In this paper, we present some aspects of a recent application of proof mining by J.Avigad, H.Towsner and the author. In this case study, we analysed a proof of the Mean Ergodic Theorem and obtained a computable rate of convergence for the ergodic averages. Proof mining generally falls into two main categories: Establishing general metatheorems that classify theorems and proofs from which additional information may be extracted and carrying out case studies. The aim of presenting aspects of a proof analysis in detail in this paper is to illustrate how the general logical results and the techniques they rely on translate into a proof analysis in practice. 1
2 1 General introduction
"... This paper (and its companion [29]) is another case study in the project of ‘proof mining ’ 1 in analysis by which we mean the logical analysis of mathematical proofs (typically using noneffective analytical tools) with the aim of extracting new numerically relevant information (e.g. effective unif ..."
Abstract
 Add to MetaCart
This paper (and its companion [29]) is another case study in the project of ‘proof mining ’ 1 in analysis by which we mean the logical analysis of mathematical proofs (typically using noneffective analytical tools) with the aim of extracting new numerically relevant information (e.g. effective uniform bounds or algorithms etc.) hidden in the proofs. 2 Many problems in numerical (functional) analysis can be seen as instances of the following general task: construct a solution x of an equation A(x): ≡ (F(x) = 0), where x is an element of some Polish (i.e. complete separable metric) space (typically with additional structure) and F: X → IR (usually F will depend on certain parameters a which again belong to Polish spaces). Quite often the construction of such a solution is obtained in two steps: 1) One shows how to construct (uniformly in the parameters of A) approximate solutions (sometimes called ‘εsolutions’) xε ∈ X for an εversion of the original equation Aε(x): ≡ ( F(x)  < ε). 2) Exploiting compactness conditions on X one concludes that either (x 1 n)n∈IN itself or some subsequence of it converges to a solution of A(x). The first step usually is constructive. However, the noneffectivity of the second step in many cases prevents one from being able to compute a solution x of A effectively within a prescribed error 1 k, i.e. to compute a function n(k) such that dX(xn(k), x) < 1 k. In many cases X: = K is compact and x is uniquely determined. Then (xn) itself converges to x so that no subsequence needs to be selected. However, the problem of how to get apriori bounds (in particular not depending on x itself) on the rate of convergence of that sequence remains. In numerical analysis, often such rates are not provided (due to the ineffectivity of the proof of the uniqueness of x). 3 In a series of papers we have demonstrated the applicability of proof theoretic techniques to extract socalled uniform moduli of uniqueness (which generalize 1 The term ‘proof mining ’ (instead of G. Kreisel’s ‘unwinding of proofs’) for the activity of extracting additional information hidden in given proofs using proof theoretic tools was suggested to the author by Professor Dana Scott. 2 For a different case study in analysis in the context of best approximation theory see [21],[22]. For other kinds of logical analyses of specific proofs see [33] and [36].