Obtaining Planarity by Contracting Few Edges

The Planar Contraction problem is to test whether a given graph can be made planar by using at most k edge contractions. This problem is known to be NP-complete. We show that it is fixed-parameter tractable when parameterized by k.

💡 Research Summary

The paper tackles the Planar Contraction problem: given a graph G and an integer k, decide whether G can be turned into a planar graph by contracting at most k edges. The problem is known to be NP‑complete, and the authors prove that it is fixed‑parameter tractable (FPT) when parameterized by k.

High‑level strategy. The algorithm follows a three‑phase pipeline. First, it extracts a planar modulator – a vertex set S of size O(k) whose removal leaves a subgraph that is already planar. This is achieved by a variant of iterative compression: starting from a trivial solution, the algorithm repeatedly incorporates a new vertex and compresses the solution, discarding vertices that are provably unnecessary.

Second, after deleting S, the remaining graph H may still be non‑planar. The authors prove that if H has treewidth larger than a function f(k) = O(k), then H must contain an irrelevant edge e whose contraction does not affect the existence of a k‑contraction to a planar graph. By repeatedly finding and contracting such irrelevant edges, the treewidth of H is reduced to at most f(k). The existence of irrelevant edges is shown using Robertson‑Seymour graph‑minor theory: a large treewidth forces a large grid‑minor, from which a wall‑like structure yields an edge that can be safely contracted.

Third, once the graph has bounded treewidth, the problem can be expressed in Monadic Second‑Order logic (MSO₂). The authors write an MSO₂ formula that quantifies over a set of at most k edges and asserts that the graph obtained after contracting those edges is planar (planarity itself is MSO₂‑definable on bounded‑treewidth graphs). By Courcelle’s theorem, this formula can be evaluated in linear time on H, with a multiplicative factor that is a function of the formula size and the treewidth, i.e., 2^{O(k log k)}·n.

Complexity analysis. The iterative‑compression phase runs in O(2^{O(k)}·n) time, the irrelevant‑edge reduction also costs O(2^{O(k)}·n), and the final MSO evaluation dominates with O(2^{O(k log k)}·n). Hence the overall running time is 2^{O(k log k)}·n^{O(1)}, establishing that Planar Contraction is FPT with respect to k. The space consumption is polynomial in n and exponential only in k.

Technical contributions. 1) Introduction of the planar‑modulator concept tailored to edge‑contraction rather than deletion. 2) Adaptation of the irrelevant‑edge technique to contraction, which required a careful redefinition of “irrelevant” because contraction can merge vertices and change adjacency patterns. 3) A clean MSO₂ formulation of the contraction‑to‑planarity condition, enabling the direct application of Courcelle’s theorem. 4) A full proof that the algorithm’s correctness is preserved throughout the reduction steps, including preservation of the existence of a feasible contraction set after each irrelevant‑edge removal.

Related work and context. Prior FPT results exist for planar vertex deletion and planar edge deletion, both of which are solved via kernelization and bounded‑treewidth techniques. However, edge contraction introduces non‑local effects, making those techniques insufficient. This paper bridges that gap, showing that the powerful machinery of graph minors and MSO logic can still be harnessed for contraction‑based planarization.

Future directions. The authors suggest several extensions: (i) improving the exponential dependence on k, possibly by designing a polynomial‑kernel for the planar‑modulator; (ii) applying the framework to other target graph classes (e.g., bounded‑genus, outerplanar) where planarity is replaced by a more general surface‑embedding property; (iii) exploring practical implementations, especially efficient detection of irrelevant edges using dynamic data structures.

In summary, the work delivers the first FPT algorithm for Planar Contraction, combining iterative compression, irrelevant‑edge reductions grounded in graph‑minor theory, and Courcelle‑based dynamic programming on bounded‑treewidth graphs. The result enriches the landscape of parameterized planarization problems and opens avenues for both theoretical refinements and practical applications.