Weakly Equilibrium Cantor Sets

• Weakly equilibrium Cantor sets are recursively defined Cantor-type compact sets with a rigid dyadic structure and a logarithmic equilibrium measure recovered as a weak-* limit of finite-level distributions.
• They exhibit a precise mass distribution where each basic interval at level s carries an equal measure of 2⁻ˢ, aligning combinatorial geometry with potential theory.
• The framework enables explicit analyses via orthogonal polynomials, dyadic recursions, and adapted Hausdorff gauge measures, linking logarithmic capacity with fractal geometry.

Searching arXiv for direct sources on “weakly equilibrium Cantor sets” and closely related follow-up work. Weakly equilibrium Cantor sets are a family of Cantor-type compact sets $K(\gamma)$ determined by a parameter sequence $\gamma=(\gamma_s)_{s=1}^\infty$, and distinguished by the fact that their logarithmic equilibrium measure is recovered as the weak-* limit of natural levelwise measures on the basic intervals of the construction. In the arXiv literature represented here, the term is attached specifically to this recursively defined family rather than to Cantor sets in general. Their study combines logarithmic potential theory, orthogonal polynomials, Hausdorff gauge measures, and dyadic recursion, and yields an unusually explicit equilibrium theory on a non-homogeneous Cantor support (Alpan et al., 2015, Alpan et al., 2014).

1. Construction of the sets $K(\gamma)$

The basic data are a sequence

$$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$

together with the recursion

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$

The associated polynomials are defined by

$$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$

and one obtains a nested family of compact sets $E_s$, each a union of $2^s$ disjoint closed basic intervals $I_{j,s}$, with

$K(\gamma)=\bigcap_{s=0}^\infty E_s.$

This yields a Cantor-type compact set with a rigid dyadic interval structure (Alpan et al., 2014).

The geometry varies sharply with $\gamma=(\gamma_s)_{s=1}^\infty$0. If $\gamma=(\gamma_s)_{s=1}^\infty$1 for all $\gamma=(\gamma_s)_{s=1}^\infty$2, then $\gamma=(\gamma_s)_{s=1}^\infty$3. If $\gamma=(\gamma_s)_{s=1}^\infty$4, then the basic interval lengths satisfy

$\gamma=(\gamma_s)_{s=1}^\infty$5

and $\gamma=(\gamma_s)_{s=1}^\infty$6 has Lebesgue measure zero. If the parameters are sufficiently close to $\gamma=(\gamma_s)_{s=1}^\infty$7, positive Lebesgue measure can occur (Alpan et al., 2015).

A central potential-theoretic invariant is the logarithmic capacity. For this family,

$\gamma=(\gamma_s)_{s=1}^\infty$8

and the non-polar regime is characterized equivalently by

$\gamma=(\gamma_s)_{s=1}^\infty$9

Different results impose different hypotheses on $K(\gamma)$0; in particular, the measure-comparison results discussed below assume the stronger summability condition

$K(\gamma)$1

(Alpan et al., 2015, Alpan et al., 2014).

2. The weak equilibrium property

The defining weak-equilibrium feature is the convergence of natural level measures to the equilibrium measure. For each level $K(\gamma)$2, let $K(\gamma)$3 be the probability measure obtained by placing mass $K(\gamma)$4 uniformly on each level-$K(\gamma)$5 basic interval $K(\gamma)$6. If $K(\gamma)$7 is nonpolar, then

$K(\gamma)$8

where $K(\gamma)$9 is the logarithmic equilibrium measure of $$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$0 (Alpan et al., 2014).

This weak-* convergence is the precise sense in which the set is called weakly equilibrium. The equilibrium measure is not introduced abstractly and then studied independently of the construction; rather, it emerges as the limit of the canonical finite-level distributions determined by the dyadic geometry. A direct consequence is the exact cylinder-mass formula

$$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$1

so every level-$$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$2 basic interval carries the same equilibrium mass (Alpan et al., 2014).

This synchronization between combinatorics and potential theory is the distinctive structural feature of the class. In more general Cantor settings, equilibrium or harmonic measure need not distribute so uniformly across the natural construction pieces. Here the dyadic decomposition, the recursive geometry, and the equilibrium measure are aligned at every level.

3. Orthogonal polynomials and dyadic recursion

Let $$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$3 denote the monic orthogonal polynomial of degree $$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$4 with respect to $$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$5. The main theorem of the orthogonal-polynomial theory is that the dyadic subsequence coincides with the Chebyshev polynomials of the set: $$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$6 Thus the degree-$$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$7 monic orthogonal polynomial is exactly the degree-$$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$8 monic Chebyshev polynomial of $$\gamma=(\gamma_s)_{s=1}^\infty,\qquad 0<\gamma_s\le \frac14,$$9 (Alpan et al., 2015).

These dyadic polynomials satisfy the nonlinear recurrence

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$0

equivalently

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$1

The corresponding norm identity is

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$2

These formulas turn the dyadic subsequence into the backbone of the entire orthogonal-polynomial system (Alpan et al., 2015).

Arbitrary degrees are organized by binary expansion. If

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$3

the associated monic product

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$4

provides an adapted basis. For

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$5

the polynomial $$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$6 lies in the span of

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$7

The first nontrivial example is

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$8

This yields a recursive, explicitly dyadic algebra for all orthogonal polynomials, not only for powers of two (Alpan et al., 2015).

4. Hausdorff gauge measure and equilibrium measure

A second major theorem identifies the equilibrium measure with a Hausdorff measure built from a gauge adapted to the construction. Assume

$$r_0=1,\qquad r_s=\gamma_s r_{s-1}^2 \quad (s\ge 1).$$9

Define

$$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$0

The level lengths satisfy

$$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$1

Now define a decreasing function $$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$2 by $$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$3, with logarithmic interpolation between successive $$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$4, and then set

$$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$5

This gives a tailored Hausdorff gauge with

$$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$6

(Alpan et al., 2014).

The resulting $$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$7-Hausdorff measure satisfies

$$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$8

so $$P_2(x)=x(x-1),\qquad P_{2^{s+1}}(x)=P_{2^s}(x)\bigl(P_{2^s}(x)+r_s\bigr),$$9 is an $E_s$0-set: $E_s$1 More importantly, under the same hypotheses and nonpolarity,

$E_s$2

The proof yields quantitative two-sided domination: $E_s$3 for Borel $E_s$4 (Alpan et al., 2014).

This result is specific to the adapted gauge $E_s$5, not to a fixed power gauge $E_s$6. The geometry is non-homogeneous, so the relevant Hausdorff measure is not generally $E_s$7. The conceptual point is that the weak-equilibrium mass rule

$E_s$8

matches the gauge normalization

$E_s$9

and the length estimates $2^s$0 transfer that levelwise agreement into global measure equivalence (Alpan et al., 2014).

The same paper proves that both $2^s$1 and $2^s$2 are regular in the Stahl–Totik sense. If $2^s$3 are the orthonormal polynomials for a compactly supported measure $2^s$4, regularity means

$2^s$5

For the present family, both the equilibrium measure and the adapted Hausdorff measure belong to $2^s$6 (Alpan et al., 2014).

5. Jacobi parameters, asymptotics, and Widom factors

Since $2^s$7 is supported on $2^s$8, the monic orthogonal polynomials satisfy the three-term recurrence

$2^s$9

with $I_{j,s}$0, $I_{j,s}$1. By symmetry of $I_{j,s}$2 about $I_{j,s}$3, the diagonal coefficients are constant: $I_{j,s}$4 Also,

$I_{j,s}$5

The off-diagonal Jacobi parameters $I_{j,s}$6 are recursively computable from the dyadic norms $I_{j,s}$7 via explicit relations derived from the binary structure of the polynomials (Alpan et al., 2015).

In the thin regime

$I_{j,s}$8

the Jacobi parameters exhibit a $I_{j,s}$9-adic asymptotic self-similarity: $K(\gamma)=\bigcap_{s=0}^\infty E_s.$0 with $K(\gamma)=\bigcap_{s=0}^\infty E_s.$1. In particular,

$K(\gamma)=\bigcap_{s=0}^\infty E_s.$2

By the cited theorem of Dombrowski, this implies that $K(\gamma)=\bigcap_{s=0}^\infty E_s.$3 has zero Lebesgue measure in that regime (Alpan et al., 2015).

The associated Widom factors are

$K(\gamma)=\bigcap_{s=0}^\infty E_s.$4

For dyadic degrees,

$K(\gamma)=\bigcap_{s=0}^\infty E_s.$5

Hence

$K(\gamma)=\bigcap_{s=0}^\infty E_s.$6

and if $K(\gamma)=\bigcap_{s=0}^\infty E_s.$7 for all $K(\gamma)=\bigcap_{s=0}^\infty E_s.$8, then

$K(\gamma)=\bigcap_{s=0}^\infty E_s.$9

Moreover,

$\gamma=(\gamma_s)_{s=1}^\infty$00

Thus the Widom factors are bounded below but unbounded above in the thin weakly equilibrium regime (Alpan et al., 2015).

These asymptotics place weakly equilibrium Cantor sets outside the standard finite-gap setting while retaining a nontrivial Widom-type lower-bound phenomenon. The papers explicitly frame this as evidence toward broader questions about equilibrium measures and Szegő-type conditions on singular Cantor supports (Alpan et al., 2015).

6. Broader context and related notions

Weakly equilibrium Cantor sets sit inside a broader equilibrium-theoretic literature on Cantor supports, but they are not synonymous with every Cantor set carrying a natural measure. A complementary framework studies a general Cantor attractor $\gamma=(\gamma_s)_{s=1}^\infty$01 of an iterated function system via finite-gap approximants

$\gamma=(\gamma_s)_{s=1}^\infty$02

their equilibrium measures $\gamma=(\gamma_s)_{s=1}^\infty$03, and discrete approximants $\gamma=(\gamma_s)_{s=1}^\infty$04, with

$\gamma=(\gamma_s)_{s=1}^\infty$05

That approach provides numerical and spectral access to equilibrium measures on Cantor IFS attractors, but it does not define the term weakly equilibrium; rather, it supplies a more general weak-approximation paradigm for equilibrium measures on fractal supports (Mantica, 2013).

Several nearby Cantor-set topics use the adjective “weak” in unrelated senses. The translation-nesting theory of “Nested Cantor sets” studies when $\gamma=(\gamma_s)_{s=1}^\infty$06 for a positive-measure set of $\gamma=(\gamma_s)_{s=1}^\infty$07, controlled by the threshold $\gamma=(\gamma_s)_{s=1}^\infty$08, but does not use the weak-equilibrium terminology (Berger et al., 2013). “Twofold Cantor sets in $\gamma=(\gamma_s)_{s=1}^\infty$09” concerns failure of the weak separation property and canonical coding of overlapping self-similar sets, again without any equilibrium-state content (Kamalutdinov et al., 2018). “Almost-nowhere intersection of Cantor sets, and sufficient sampling of their cumulative distribution functions” studies deleted-digit Cantor sets with their canonical Hutchinson invariant measures and CDFs, not weakly equilibrium sets in Goncharov’s sense (Byars et al., 2019).

This suggests a useful terminological distinction. Weakly equilibrium Cantor sets are not simply Cantor sets with weak separation, weak nesting stability, or a canonical self-similar invariant measure. In the literature represented here, they are a specific recursive family $\gamma=(\gamma_s)_{s=1}^\infty$10 for which the equilibrium measure is levelwise approximable, dyadically distributed on the basic intervals, compatible with an adapted Hausdorff gauge, and sufficiently rigid to support an explicit orthogonal-polynomial theory (Alpan et al., 2015, Alpan et al., 2014).