Subsets of Ahlfors-regular curves and Jones numbers (Q677854)

The author proves that an Ahlfors-regular set (with dimension one) \(E \subseteq R^n\) which satisfies a \(\beta_q\)-version of P. W. Jones' geometric lemma is included in an Ahlfors-regular curve \(\Gamma\). A subset \(E \subseteq R^n\) is Ahlfors-regular of dimension 1 if and only if \(E\) is closed and if there is a \(C_0>0\) such that for every \(x \in E\), for every \(R \in (0, \text{diam } E)\), \[ C_0^{-1} R \leq H^1(E \cap B(x,R)) \leq C_0 R, \] where \(H^1\) denotes the Hausdorff 1 measure. The numbers \(\beta_{\infty}\) of Jones are defined for every \(x \in R^n\), for every \(t>0\) by, if \(E \cap B(x,t) \neq \emptyset\), \[ \beta_{\infty}(x,t,E) = \inf_L \sup_{y \in E \cap B(x,t)} \left(\frac{\text{dist}(y,L)}{t} \right) \] and the inf is taken over all lines in \(R^n\), and when \(E \cap B(x,t) = \emptyset\), \(\beta_{\infty}(x,t,E) = 0\). \(\beta_{\infty}\) measures in every ball how well \(E\) is approximated by lines. Jones used them to give a characterization of subsets of rectifiable curves. For every \(q \geq 1\), every \(x \in E\), every \(t>0\), define \[ \beta_q(x,t,E) = \inf_L \left( \frac 1t \int_{y \in E \cap B(x,t)} \left(\frac{\text{dist}(y,L)}{t} \right)^q dH^1(y) \right)^{1/q} \] where again the inf is taken over all lines in \(R^n\). The author's main result is Theorem 1.2. Let \(E\) be a compact Ahlfors-regular set of dimension 1 in \(R^n\). If \(\beta_q(x,t,E)^2 dH^1(x) \frac{dt}{t}\) defines a Carleson measure on \(E \times R^+\), i.e., there is a constant \(C>0\) such that for every \(x \in E\), every \(R>0\), \[ \int_{y \in E \cap B(x,R)} \int_0^R \beta_q(y,t,E)^2 dH^1(y) \frac{dt}{t} \leq CR, \] then \(E \subseteq \Gamma\), where \(\Gamma\) is an Ahlfors-regular curve. The theorem is due to \textit{G. David} and \textit{S. Semmes} [Astérisque 193, 145 p. (1991; Zbl 0743.49018)]; the author gives a more direct proof inspired by the ``insertion of the nearest neighbor'' algorithm used for the travelling salesman problem. The latter technique had been used by \textit{P. W. Jones} [Invent. Math. 102, No. 1, 1-15 (1990; Zbl 0731.30018)] in his proof of the \(q = \infty\) case of the theorem; Jones' proof would not extend to \(q \neq \infty\); the construction of David and Semmes is more complicated than the construction of the author.