Two-point pre-image entropy (Q2642812)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Two-point pre-image entropy |
scientific article |
Statements
Two-point pre-image entropy (English)
0 references
5 September 2007
0 references
Two new entropy-like invariants are introduced in this paper. They concern continuous endomorphisms \(T\) from a compact metric topological group \((X,d,\cdot)\) into itself and are defined as follows, where \([x]=\{x,x^{-1}\}\), \([\xi]\) is the \(\sigma\)-algebra generated by the partition \(\{[x]: x\in X\}\), and \([\xi]^-=\bigcap_{k\geq 0} T^{-k}[\xi]\). Let \(r(n,\varepsilon,T,A)\) denote the maximum cardinality of an \((n,\varepsilon)\)-separating set of \(T\) for \(A\). Then the two-point pre-image entropy of \(T\) is defined by \[ h_{\text{top}}(T| [\xi]^-)=\lim_{\varepsilon\rightarrow 0} \limsup_{n\rightarrow \infty} \frac{1}{n} \log \sup_{k\geq 0, x\in X} r(n,\varepsilon,T, T^{-k}([x]). \] Now assume that \(\mu\) is an invariant Borel probability measure under \(T\), write \(\alpha^n=\bigvee_{i=0}^n T^{-i}\alpha\) for every finite partition \(\alpha\) of \(X\), and let \(H_\mu(\beta| \mathcal{C})\) denote the conditional entropy of a finite partition \(\beta\) given a sub-\(\sigma\)-algebra \(\mathcal{C}\) of the Borel \(\sigma\)-algebra of \(X\). Then the two-point conditional entropy of \(T\) with respect to \(\mu\) and \([\xi]^-\) is given by \[ h_\mu(T| [\xi]^-,\alpha)=\lim_{n\rightarrow \infty} \frac{1}{n} H_\mu(\alpha^n| [\xi]^-) \] and \(h_\mu(T| [\xi]^-)=\sup_\alpha h_\mu(T| [\xi]^-,\alpha)\). After showing that two-point entropies \(h_{\text{top}}(T| [\xi]^-)\) and \(h_\mu(T| [\xi]^-)\) satisfy some basic properties of standard topological and metric entropy (power rules, product rules and affineness), the main result of the paper (the variational principle) is proved: \(h_{\text{top}}(T| [\xi]^-)=\sup_{\mu\in M(X,T)} h_\mu(T| [\xi]^-)\), where \(M(X,T)\) is the collection of all invariant measures \(\mu\) under \(T\).
0 references
conditional entropy
0 references
preimage entropy
0 references
variational principle
0 references