[Paper Review] On the Assouad dimension and convergence of metric spaces
This paper introduces pseudo-cones as a generalization of tangent and asymptotic cones in metric spaces, proving that the Assouad dimension of any metric space is bounded below by that of its pseudo-cones. It establishes that the Assouad dimension is stable under Gromov–Hausdorff limits and ultralimits, and constructs an $(\omega_0+1)$-metric space containing all compact metric spaces as pseudo-cones, demonstrating that analogies of such dimension bounds fail for topological, Hausdorff, and conformal Hausdorff dimensions.
We introduce the notion of pseudo-cones of metric spaces as a generalization of both of the tangent cones and the asymptotic cones. We prove that the Assouad dimension of a metric space is bounded from below by that of any pseudo-cone of it. We exhibit a example containing all compact metric spaces as pseudo-cones, and examples containing all proper length spaces as tangent cones or asymptotic cones.
Motivation & Objective
- To generalize the concept of tangent and asymptotic cones via pseudo-cones for studying metric space dimensions.
- To establish lower bounds for the Assouad dimension of a metric space using its pseudo-cones.
- To demonstrate that the Assouad dimension is preserved under Gromov–Hausdorff convergence and ultralimit constructions.
- To construct a single metric space whose pseudo-cones include all compact metric spaces, and another whose tangent or asymptotic cones include all proper length spaces.
- To show that analogies of Assouad dimension bounds do not hold for topological, Hausdorff, or conformal Hausdorff dimensions.
Proposed method
- Define pseudo-cones as Gromov–Hausdorff limits of scaled subsets $ u_i A_i $ of a metric space $ X $, where $ u_i \in (0,\infty) $.
- Use the stability of the Assouad dimension under metric scaling: $ \dim_A(hX) = \dim_A(X) $ for $ h > 0 $.
- Apply techniques from Mackay and Tyson (2010) on tangent spaces to extend lower bounds on Assouad dimension to pseudo-cones.
- Construct a metric space $ X $ using nested sets $ G_i $, with controlled distances and scaling factors $ a_i $, to ensure desired convergence properties.
- Utilize ultralimits via non-principal ultrafilters to analyze limit behavior of scaled subsets, proving dimension bounds for ultralimit pseudo-cones.
- Employ combinatorial and geometric estimates (e.g., $ \alpha(G_i)/\alpha(G_{i-1}) < 16 $) to control distances and convergence rates in the construction.
Experimental results
Research questions
- RQ1Can the notion of tangent and asymptotic cones be generalized to a broader class of limit objects in metric spaces?
- RQ2Does the Assouad dimension of a metric space bound the Assouad dimension of its pseudo-cones?
- RQ3Can a single metric space be constructed such that all compact metric spaces appear as pseudo-cones?
- RQ4Do dimension bounds based on tangent or asymptotic cones extend to conformal Assouad dimension?
- RQ5Are analogies of Assouad dimension bounds valid for topological, Hausdorff, or conformal Hausdorff dimensions?
Key findings
- For any metric space $ X $, the Assouad dimension of every pseudo-cone $ P \in \mathrm{PC}(X) $ satisfies $ \dim_A P \leq \dim_A X $.
- The Assouad dimension of any ultralimit $ \lim_{\mathcal{U}}(u_i A_i, a_i) $ is bounded above by $ \dim_A X $.
- There exists an $ (\omega_0+1) $-metric space $ X $ such that $ \mathrm{PC}(X) $ contains all compact metric spaces as pseudo-cones.
- There exists an $ (\omega_0+1) $-metric space $ X $ for which every proper length space appears as a tangent or asymptotic cone.
- The conformal Assouad dimension satisfies $ \mathrm{Cdim}_A P \leq \mathrm{Cdim}_A X $ for every $ P \in \mathrm{KPC}(X) $, the class of compact pseudo-cones.
- The topological, Hausdorff, and conformal Hausdorff dimensions do not satisfy analogous lower bounds: there exist $ X $ and $ P \in \mathrm{PC}(X) $ such that $ \dim_T X < \dim_T P $, $ \dim_H X < \dim_H P $, and $ \mathrm{Cdim}_H X < \mathrm{Cdim}_H P $.
Better researchstarts right now
From reading papers to final review, dramatically reduce your research time.
No credit card · Free plan available
This review was created by AI and reviewed by human editors.