Entropy of flows, revisited (Q4950753)

Let \(\Phi\) denote a flow on a compact metric space \(M\) and let \(\mu\) be a \(\Phi\)-invariant measure. Let \(N( \delta,t, \varepsilon)\) denote the minimal number of sets \(B\) needed to cover a set of \(\mu\)-measure at least \(\delta\), where the sets \(B\) are of the form \[ \{y\in M: \exists\text{ increasing continuous }\alpha:[0,t]\to [0,\alpha(t)];\;\sup_{s\leq t}\operatorname { dist} (\Phi_{\alpha(s)} (x), \Phi_s(y))< \varepsilon\}. \] The entropy of \(\mu\) is defined by \[ e_\mu(\Phi)= \lim_{\varepsilon\to 0} \limsup_{t\to \infty} t^{-1} \log N(\delta, t, \varepsilon). \] (The limit does not depend on \(\delta\).) The topological entropy is defined by its supremum over all invariant measures. Invariance properties of this notion are derived in this article. Another result confirms a problem of \textit{R. Bowen} and \textit{P. Walters} [J. Differ. Equ. 12, 180-193 (1972; Zbl 0242.54041)] that the topological entropy of the time-one-map of an expansive flow equals that of its symbolic representation.