S-Measure and Minkowski Equivalence
- S-measure is a surface-area based metric that defines the s-dimensional content of a bounded set via the asymptotic behavior of its parallel neighborhoods.
- It demonstrates that S-measurability and Minkowski measurability coincide, with both approaches yielding identical limits and two-sided bounds.
- The framework extends to generalized gauge functions and one-dimensional fractal strings, bridging geometric content with spectral asymptotics.
Searching arXiv for the specified paper to ground the article in the cited source. S-measure, in the sense developed through the surface-area based content of parallel sets, is the notion of measuring a bounded set through the asymptotic behavior of the surface area of the boundary of its -parallel neighborhood as . In the formulation of Rataj and Winter, the relevant object is the -dimensional S-content, defined in direct analogy with the classical Minkowski content, and a set is called -S-measurable when this content exists and is positive and finite. The central result is that Minkowski measurability and S-measurability coincide, with exact agreement of the corresponding limits, and that this equivalence extends to two-sided positivity and finiteness bounds as well as to generalized contents defined by gauge functions (Rataj et al., 2011).
1. Parallel sets and the classical Minkowski content
Let be bounded. The Euclidean distance from to is
For , the 0-parallel neighborhood is
1
Its 2-dimensional volume is
3
Fix 4. The 5-dimensional lower and upper Minkowski contents of 6 are defined by
7
where
8
is the volume of the unit 9-ball. Whenever
0
their common value 1 is called the 2-dimensional Minkowski content of 3, and 4 is said to be 5-Minkowski measurable (Rataj et al., 2011).
This framework measures the small-scale growth of the volume of parallel sets. In the terminology used by Rataj and Winter, the S-content is introduced in complete analogy with this classical volume-based construction.
2. Definition of S-content and S-measurability
The surface-area analogue is obtained by setting
6
the surface area of the boundary of 7. The lower and upper 8-dimensional S-contents are then defined by
9
If
0
their common value 1 is called the 2-dimensional S-content of 3, and 4 is said to be 5-S-measurable (Rataj et al., 2011).
The normalization is chosen so that the asymptotic scaling of 6 matches that of the derivative of 7. This suggests that S-content is not merely analogous to Minkowski content but is structurally linked to it through the geometry of parallel sets.
3. Equivalence with Minkowski measurability
The central theorem states that Minkowski measurability and S-measurability coincide. Let 8 be bounded, let 9, and set 0. Then the following are equivalent:
- 1 is 2-Minkowski measurable, that is,
3
- 4 is 5-S-measurable, that is,
6
Moreover, in that case one has the exact coincidence of the two limits: 7 In particular, the common exponent 8 is then both the Minkowski and the S-dimension of 9 (Rataj et al., 2011).
A further theorem gives the corresponding two-sided criterion without requiring existence of a limit. For bounded 0 and fixed 1,
2
if and only if
3
Thus any two-sided positive finite bound for the Minkowski content at exponent 4 forces the same two-sided bound for the S-content, and vice versa (Rataj et al., 2011).
These results characterize S-measure as a fully equivalent surface-area formulation of Minkowski measurability rather than a weaker proxy. A plausible implication is that, for bounded sets in Euclidean space, the asymptotic geometry of parallel volume and parallel surface area carries the same measurable information at the critical exponent.
4. Kneser functions and the analytic mechanism
The equivalence is derived from analogous statements for Kneser functions. A function 5 is a Kneser function of order 6 if for all 7 and 8 one has
9
Rataj and Winter establish two general principles for such functions. First, if 0 is Kneser of order 1 and
2
then also
3
and conversely. Second, if in addition the limit
4
exists, then
5
also exists and equals 6 (Rataj et al., 2011).
These principles are applied to the volume function 7, whose left and right derivatives satisfy, in the sense of measures, 8. This immediately yields the corresponding statements for Minkowski and S-contents. The role of the Kneser-function framework is therefore foundational: it transfers asymptotic information between a function and its derivative, and in the geometric setting those two objects are the volume and surface area of parallel sets.
5. Generalized contents and gauge functions
The theory extends beyond power-law normalizations. Let 9 be a continuous gauge function with 0 as 1. The generalized Minkowski contents are defined by
2
and similarly for the generalized S-contents with the same 3.
Under mild regularity assumptions on 4, one again has an equivalence theorem: if 5 is differentiable near 6, 7 does not vanish, and the generalized Minkowski content 8 exists in 9, then the generalized S-content 0 exists in 1 and equals 2, and conversely (Rataj et al., 2011).
In particular, for
3
with 4 nondecreasing and slowly varying, one has
5
and conversely. This shows that the relation between Minkowski content and S-content is not restricted to pure dimensional scaling. It persists for more general gauges, provided the gauge has sufficient differentiability and nondegeneracy near the origin.
6. One-dimensional fractal strings and the Modified Weyl–Berry conjecture
In dimension 6, the theory specializes to compact sets 7 of Minkowski dimension 8, studied via their complementary intervals, or fractal string, 9. Lapidus–Pomerance showed that 0 is Minkowski measurable of dimension 1 if and only if
2
and used this to resolve the modified Weyl–Berry conjecture for planar vibrations (Rataj et al., 2011).
Rataj and Winter add a further equivalent criterion. For compact 3 with 4, the following are all equivalent:
- 5,
- 6,
- 7 as 8.
Moreover, exact measurability satisfies
9
and in this case
00
By passing through the S-content, several steps in the original Lapidus–Pomerance proofs can be streamlined, in particular the sharp two-sided estimates of the counting error for eigenvalues of the Dirichlet Laplacian on the complement of 01 (Rataj et al., 2011).
This one-dimensional application situates S-measure within spectral asymptotics and fractal-string theory. It also shows that the surface-area formulation is not only equivalent at an abstract level but can serve as an effective intermediate criterion in problems where interval asymptotics and geometric content are linked.