Well-Formed Separator Sequences, with an Application to Hypergraph Drawing

Given a hypergraph $H$, the Planar Support problem asks whether there is a planar graph $G$ on the same vertex set as $H$ such that each hyperedge induces a connected subgraph of $G$. Planar Support is motivated by applications in graph drawing and data visualization. We show that Planar Support is fixed-parameter tractable when parameterized by the number of hyperedges in the input hypergraph and the outerplanarity number of the sought planar graph. To this end, we develop novel structural results for $r$-outerplanar triangulated disks, showing that they admit sequences of separators with structural properties enabling data reduction. This allows us to obtain a problem kernel for Planar Support, thus showing its fixed-parameter tractability.

💡 Research Summary

The paper tackles the Planar Support problem, which asks whether a given hypergraph H = (V, E) admits a planar graph G on the same vertex set such that every hyperedge induces a connected subgraph of G. This problem is central to hypergraph drawing, subdivision drawings, and various visualization tasks. While the problem is known to be NP‑complete in general, its parameterized complexity has remained largely unexplored, especially when the desired support must be r‑outerplanar (i.e., it can be embedded with at most r layers of nesting).

The authors’ main contribution is to prove that Planar Support is fixed‑parameter tractable (FPT) when parameterized by the combined parameters m = |E| (the number of hyperedges) and r (the outerplanarity bound of the sought support). They achieve this by developing a novel structural tool: well‑formed separator sequences for r‑outerplanar triangulated disks. A triangulated disk is a planar graph whose interior faces are all triangles and whose outer face is a simple cycle; when such a graph is r‑outerplanar, its vertices can be partitioned into layers L₁,…,L_r based on distance from the outer face.

A well‑formed separator sequence is a list of triples (A_i, S_i, B_i) where each separator S_i has a fixed width p, consists either of an induced path or an induced cycle, and satisfies a series of stringent properties:

1. Linear separation (A_i ∪ B_i = V, A_i ∩ B_i = S_i, no edges cross between A_i \ S_i and B_i \ S_i, and the triples are nested).
2. Simple shape (all separators are uniformly paths or cycles, with endpoints on the outer layer).
3. Uniqueness of vertex positions (the k‑th vertex on any separator lies on a unique layer, and the same index across different separators always refers to vertices on the same layer).
4. Layering constraints (each separator uses at most two vertices from any layer).

The authors prove (Theorem 3.2) that any r‑outerplanar triangulated disk with n vertices contains a well‑formed separator sequence of length at least Ω(r·log n) and width at most 2r. This construction is algorithmic and runs in polynomial time.

A crucial operation introduced is gluing: given two separators S_i and S_j (i < j) from the sequence, one removes everything between them, keeps the subgraphs induced by A_i and B_j, and identifies the corresponding vertices of S_i and S_j (preserving their order). Lemma 3.4 shows that the resulting graph remains r‑outerplanar. This property enables the authors to “compress” large portions of a candidate support without destroying the outerplanarity or the connectivity requirements of the hyperedges.

The next step is to translate these graph‑theoretic insights into a kernelization for Planar Support. The authors define separator signatures, which capture how hyperedges intersect a separator. If the hypergraph is large relative to the parameters, the existence of two separators with identical signatures implies that the region between them is redundant for the connectivity of all hyperedges. By removing this region and gluing the two separators, one obtains a smaller instance that is equivalent with respect to the existence of an r‑outerplanar support. Repeating this reduction yields a problem kernel whose size is bounded by a function of m and r (specifically O(m·r·log n)). Consequently, a deterministic algorithm runs in time f(m, r)·poly(n), establishing the FPT result.

An additional, subtle contribution concerns twins—pairs of vertices that belong to exactly the same set of hyperedges. Prior work often assumed twin‑free hypergraphs, arguing that twins do not affect planarity. The authors demonstrate that twins can be essential: removing twins may destroy all planar supports, while keeping them may allow a support. Their kernelization explicitly handles twins, showing that a bounded number of twins can be safely removed without altering the answer, but the presence of many twins can be leveraged to achieve the kernel bound.

Overall, the paper delivers four major advances:

1. Introduction of well‑formed separator sequences for r‑outerplanar triangulated disks, with provable length and width guarantees.
2. Proof that gluing any two separators from such a sequence preserves r‑outerplanarity.
3. A kernelization framework for Planar Support based on separator signatures and gluing, yielding a kernel of size polynomial in m and r.
4. A thorough analysis of twins, clarifying their impact on the problem’s complexity and integrating them into the FPT algorithm.

These results not only settle the fixed‑parameter tractability of Planar Support for the natural combined parameter (hyperedge count, outerplanarity) but also provide new combinatorial tools that may be applicable to other planar‑graph‑based problems, such as graph drawing, network design, and parameterized graph separation. The work bridges a gap between structural graph theory and algorithmic hypergraph visualization, offering both theoretical insight and practical algorithmic implications.