Clustered Planarity
Definition
A graph is cluster planar if there is a planar drawing of it where we can also add its clusters as closed regions that only cross the edges that leave the respective cluster.
Formally, given a graph \(G=(V,E)\) equipped with a cluster hierarchy \(T\), which is a rooted tree with the vertices \(V\) as leaves. Each inner node \(\mu\) of \(T\) represents a cluster encompassing all leaves \(V_\mu\) of the subtree rooted at \(\mu\). A cluster planar drawing is a planar drawing that also maps every cluster \(\mu\) to a simple closed region \(R_\mu\) such that
- \(R_\mu\) encloses exactly the vertices \(V_\mu\),
- no two cluster region boundaries intersect, and
- no edge intersects the boundary of a cluster more than once. A cluster graph is cluster planar if it admits a cluster planar drawing.
Background
Lengauer (1989) studied and solved this problem as early as 1989 in the setting where the clusters are connected. Feng, Cohen, and Eades (1995), who coined the term Clustered Planarity, rediscovered this algorithm and asked the general question where disconnected clusters are allowed. This question remained open for 30 years. In that time, polynomial-time algorithms were found for many special-cases (Angelini and Da Lozzo 2019; Cortese et al. 2008; Fulek et al. 2015; Gutwenger et al. 2002) before Fulek and Tóth (2022) found an \(O((n+d)^8)\) solution in 2019, where \(d\) is the number of crossings between a cluster-border and an edge leaving the cluster. Subsequently, Bläsius, Fink, and Rutter (2023) gave a solution via Synchronized Planarity with a running time of \(O((n+d)^2)\), which also reveals the main concepts for solving Clustered Planarity.
Complexity
Let \(d\) be the number of edge-cluster boundary crossings and \(\Delta\) be the maximum number of edges crossing a single cluster border. While there are no direct solutions to the problem, Clustered Planarity can be solved in time…
- \(O((n+d)^2)\subset O(n^4)\) via Synchronized Planarity (Bläsius, Fink, and Rutter 2023)
- \(O(n+d\cdot \Delta)\subset O(n^4)\) via Synchronized Planarity with SPQR-node expansion (Fink and Rutter 2023)
- \(O((n+d)^8)\subset O(n^{16})\) via Atomic Embeddability (Fulek and Tóth 2022)
All these reductions use the cd-Tree representation of Clustered Planarity (Bläsius and Rutter 2016). As the reduction subdivides an edge for each cluster boundary it crosses, the representation has size \(O(n+d)\).
Clustered Planarity also reduces to SEFE-2 by representing the cluster boundary, where the order of edges leaving needs to correspond to the order of the same edges entering, via a gadget (Schaefer 2013, Theorem 6.17; see also Rutter 2020, sec. 13.2.2).
\(\langle \alpha, \beta, \gamma \rangle\)-drawings are a generalization of Clustered Planarity, where given counts of edge-edge, cluster-cluster, and cluster-edge crossings are allowed (Angelini et al. 2015).
See also
- Wikipedia Article
- Handbook on Graph Drawing and Visualization (Patrignani 2013, sec. 1.8.2)