Annular Non-Crossing Double-Line Diagrams

• Non-crossing double-line diagrams are annular configurations where paired arcs along concentric circles form non-intersecting matchings that preserve the underlying combinatorics.
• The methodology reduces double-line ribbon graphs to planar non-crossing matchings via contraction and uses Burnside’s Lemma to derive closed-form Catalan-type enumeration formulas.
• Bijections with planar trees, binary necklaces, and applications to affine Coxeter groups highlight practical insights including generating functions, asymptotic growth, and lattice structure.

Non-crossing double-line diagrams are annular planar configurations in which ribbon-like, or “double-line,” edges can be reduced to non-intersecting arcs without losing the underlying combinatorics. In the annular framework of "Annular Non-Crossing Matchings" (Drube et al., 2015), each double-line edge is represented by two parallel arcs in the annulus, one on the outer boundary and one on the inner boundary, joined by short cross-cuts; contracting each such pair yields an annular non-crossing matching, while fattening each matching arc back to a narrow ribbon recovers a double-line graph. A distinct but related annular formalism, developed for affine Coxeter groups of types $\tilde B$ and $\tilde D$, studies symmetric noncrossing partitions of an annulus with one or two double points and uses planar annular diagrams to model intervals in absolute order (Reading, 2023).

1. Annular matchings as the combinatorial core

An annular non-crossing matching is defined on two concentric circles in the plane, denoted $S^1_{\rm out}$ and $S^1_{\rm in}$. One chooses $n$ distinct exterior points on $S^1_{\rm out}$ and $m$ distinct interior points on $S^1_{\rm in}$, with $n+m$ even, and considers a collection of $\tfrac{n+m}{2}$ smooth, non-intersecting arcs in the annular region joining the marked boundary points in pairs. An arc joining two points on $\tilde D$0 is an external half-circle; an arc joining two points on $\tilde D$1 is an internal half-circle; and an arc joining one boundary to the other is a cross-cut. If there are exactly $\tilde D$2 cross-cuts, the matching lies in $\tilde D$3, and

$\tilde D$4

Two matchings are equivalent if they are ambiently isotopic in the annulus, with endpoints allowed to slide along the same boundary circle, but without rotation of an entire boundary or passage of arcs through the hole; reflections that cannot be realized by isotopy are distinct (Drube et al., 2015).

This definition generalizes both linear non-crossing matchings in the half-plane and circular non-crossing matchings. The annulus introduces a second boundary component and therefore a genuine distinction between boundary-preserving half-circles and boundary-bridging cross-cuts. A common simplification is to regard annular non-crossing objects as merely disk diagrams with a hole removed; the annular definition is stricter, because isotopy is taken relative to the boundary structure and the hole cannot be crossed.

2. Double-line interpretation and planarity on the annulus

In physics, double-line diagrams are often drawn on an annulus to encode perturbative expansions in matrix models. In this representation, each double-line edge may be thought of as two parallel arcs in the annulus, one on the outer boundary and one on the inner boundary, joined by short cross-cuts. Contracting each pair of parallel arcs to a single arc produces exactly an annular non-crossing matching. Conversely, given a matching in $\tilde D$5, one may fatten each arc to a narrow ribbon and regard it as the projection of a double-line graph on the annulus. Under this correspondence, non-crossing is equivalent to the usual planarity of the ribbon graph on the annular surface (Drube et al., 2015).

This correspondence isolates the combinatorial skeleton of annular ribbon graphs. It also clarifies that “non-crossing” in the double-line setting is not a statement about individual centerlines alone, but about the planar embeddability of the associated ribbon structure. A plausible implication is that enumeration of annular non-crossing matchings provides a direct counting framework for annular sectors of double-line expansions in which planarity is the primary constraint.

3. Exact enumeration and the Catalan-type regime

The counting problem for annular non-crossing matchings admits a closed form obtained via Burnside’s Lemma. After separating the cases $\tilde D$6 and $\tilde D$7 and summing over the number of cross-cuts, one obtains the compact formula

$\tilde D$8

where $\tilde D$9 is Euler’s totient. The enumeration is symmetric,

$S^1_{\rm out}$0

and satisfies the parity constraint $S^1_{\rm out}$1 unless $S^1_{\rm out}$2 is even. When one boundary carries no cross-cuts,

$S^1_{\rm out}$3

In the one-boundary specialization,

$S^1_{\rm out}$4

so in particular $S^1_{\rm out}$5, where $S^1_{\rm out}$6 (Drube et al., 2015).

These formulas realize $S^1_{\rm out}$7 as a one-parameter generalization of Catalan-type enumeration. Small values reflect the same parity structure and symmetry: for example, $S^1_{\rm out}$8, $S^1_{\rm out}$9, $S^1_{\rm in}$0, and $S^1_{\rm in}$1. The factorization for $S^1_{\rm in}$2 shows that the genuinely annular interaction arises from cross-cuts; when cross-cuts are absent, the two boundaries decouple into independent Catalan counts.

4. Generating functions and asymptotic growth

The bivariate ordinary generating function is

$S^1_{\rm in}$3

A closed form is

$S^1_{\rm in}$4

where

$S^1_{\rm in}$5

The intermediate function $S^1_{\rm in}$6 is the generating function of the sequence $S^1_{\rm in}$7. The diagonal generating function $S^1_{\rm in}$8 is obtained by setting $S^1_{\rm in}$9 (Drube et al., 2015).

The leading $n$0 term of the Burnside sum governs the asymptotics. For large $n$1,

$n$2

For the total number

$n$3

one has

$n$4

Combinatorially, this matches the usual $n$5 growth of non-crossing matchings, now distributed across all boundary partitions $n$6. A simple linear recurrence can be extracted from the generating function via binomial expansions in the Burnside sum, but no compact small-order linear recurrence is known.

5. Bijections with planar graphs and binary necklaces

Several subclasses of annular non-crossing matchings admit direct bijections to other combinatorial objects. The class $n$7 is exactly the class of unrooted planar trees with $n$8 vertices and one distinguished root-region. More generally, for $n$9, the set $S^1_{\rm out}$0 is in bijection with connected planar graphs having exactly one simple $S^1_{\rm out}$1-cycle, together with $S^1_{\rm out}$2 chords inside and $S^1_{\rm out}$3 chords outside that cycle. In the maximal-cross-cut case $S^1_{\rm out}$4,

$S^1_{\rm out}$5

where $S^1_{\rm out}$6 is the number of binary necklaces with $S^1_{\rm out}$7 black beads and $S^1_{\rm out}$8 white beads. The bijection sends each white endpoint of an external half-circle to a white bead and each black endpoint, either of an external half-circle or of a cross-cut, to a black bead; rotational equivalence of necklaces matches the rotational isotopy of the annular matching (Drube et al., 2015).

These bijections locate annular double-line combinatorics within familiar enumerative regimes. The planar-graph interpretation makes the role of cross-cuts especially transparent: they generate the unique simple cycle that separates the inside and outside chord systems. The necklace interpretation packages a highly constrained annular boundary pattern into a cyclic word, which is often more convenient for closed counting formulas.

6. Symmetric annular diagrams with double points

A related annular theory replaces pairwise matchings by symmetric noncrossing partitions on an oriented topological annulus $S^1_{\rm out}$9 with one or two distinguished double points. In type $m$0, one fixes $m$1 and places $m$2 equally spaced numbered points on each boundary together with a single double point $m$3 carrying labels $m$4. In type $m$5, one fixes $m$6, places $m$7 numbered points on each boundary, and introduces two double points on each boundary. The annulus is equipped with the involution $m$8 given by rotation through $m$9. A symmetric noncrossing partition $S^1_{\rm in}$0 is a collection of disjoint embedded blocks, each of which is a trivial block, a curve block, a disk block, or an annular block; at most two annular blocks are allowed, every numbered point lies in some block, and $S^1_{\rm in}$1 permutes the blocks. The noncrossing condition forbids intersections of arcs, or of their $S^1_{\rm in}$2-images, in the interior of $S^1_{\rm in}$3 (Reading, 2023).

The resulting poset is graded of rank $S^1_{\rm in}$4, with rank function

$S^1_{\rm in}$5

For affine Coxeter groups of type $S^1_{\rm in}$6 and $S^1_{\rm in}$7, the interval $S^1_{\rm in}$8 in absolute order is isomorphic to the subposet of symmetric noncrossing partitions with no dangling annular blocks, while the larger interval $S^1_{\rm in}$9 corresponds to all symmetric noncrossing partitions. The map $n+m$0 reads off a permutation by traversing the boundaries of disk-blocks, arc-blocks, and symmetric annular blocks, and by recording the tiny cycle $n+m$1 when a double point lies in the interior. Enumeration is also explicit: if $n+m$2 counts type-$n+m$3 partitions, then

$n+m$4

with generating function $n+m$5 and closed form

$n+m$6

If $n+m$7 counts type-$n+m$8 partitions, then

$n+m$9

with generating function $\tfrac{n+m}{2}$0 and closed form

$\tfrac{n+m}{2}$1

Alternatively, $\tfrac{n+m}{2}$2 counts plane ternary trees with $\tfrac{n+m}{2}$3 leaves. The lattice-theoretic refinement adds factored translations to restore the lattice property: in affine types $\tfrac{n+m}{2}$4 and $\tfrac{n+m}{2}$5, the interval $\tfrac{n+m}{2}$6 fails to be a lattice only because certain translations have length $\tfrac{n+m}{2}$7 and must be factored, and adjoining the factored translations yields a finite lattice $\tfrac{n+m}{2}$8 (Reading, 2023).

This second theory is not identical to annular non-crossing matchings, but it belongs to the same annular planar lineage. The matching model encodes pairings of boundary points and their double-line realizations; the double-point model encodes symmetric block decompositions, affine permutations, and Garside-type lattice structure. Together they show that annular noncrossing geometry supports both exact enumerative formulas and algebraic realizations of affine Coxeter combinatorics.