Planar Graphs
A planar graph is a graph that can be drawn in the plane without any edge crossings. Planarity has deep connections to topology, graph coloring, and algorithm design.
Planar Embeddings
A planar embedding (or planar drawing) of a graph is a mapping of vertices to points and edges to curves in the plane such that edges only intersect at shared endpoints.
A graph is planar if it admits at least one planar embedding. A non-planar graph has no crossing-free drawing.
A face is a connected region of the plane bounded by edges in a planar embedding. Every planar embedding has exactly one unbounded (outer) face.
Example: K₄ (tetrahedron) is planar with 4 faces (3 triangular inner faces + 1 outer face).
Euler's Formula
For a connected planar graph with v vertices, e edges, and f faces:
v - e + f = 2
This is one of the most beautiful results in mathematics, connecting topology to combinatorics.
Example: K₄ has v = 4, e = 6, f = 4. Check: 4 - 6 + 4 = 2. ✓
Generalization: For a planar graph with c connected components:
v - e + f = 1 + c
Consequences of Euler's Formula
For simple connected planar graphs with v ≥ 3:
e ≤ 3v - 6
Proof: Each face is bounded by at least 3 edges, and each edge borders at most 2 faces. So 2e ≥ 3f, giving f ≤ 2e/3. By Euler: v - e + f = 2, so f = 2 - v + e. Then 2 - v + e ≤ 2e/3, which gives e ≤ 3v - 6. ∎
Corollary: K₅ is not planar. (v = 5, e = 10, but 3(5) - 6 = 9 < 10.)
For triangle-free planar graphs (no cycles of length 3):
e ≤ 2v - 4
Proof: Each face has at least 4 boundary edges. So 2e ≥ 4f, giving f ≤ e/2. Then e ≤ 2v - 4. ∎
Corollary: K₃,₃ is not planar. (v = 6, e = 9, bipartite so triangle-free, but 2(6) - 4 = 8 < 9.)
Minimum degree bound: Every simple planar graph has a vertex of degree ≤ 5.
Proof: If every vertex has degree ≥ 6, then 2e = Σ deg(v) ≥ 6v, so e ≥ 3v. But e ≤ 3v - 6 < 3v. Contradiction. ∎
Kuratowski's Theorem
Theorem (Kuratowski, 1930): A graph is planar if and only if it does not contain a subdivision of K₅ or K₃,₃ as a subgraph.
A subdivision of G replaces edges with paths (inserting degree-2 vertices).
K₅: Complete graph on 5 vertices (non-planar)
K₃,₃: Complete bipartite 3+3 (non-planar)
These are the two fundamental obstructions to planarity.
Testing planarity: This can be done in O(V + E) time using algorithms by Hopcroft-Tarjan (1974) or Boyer-Myrvold (2004).
Wagner's Theorem
Theorem (Wagner, 1937): A graph is planar if and only if it does not contain K₅ or K₃,₃ as a minor.
A graph minor of G is obtained by:
- Deleting vertices
- Deleting edges
- Contracting edges (merging endpoints)
Wagner's theorem is equivalent to Kuratowski's but uses minors instead of subdivisions. Minors are often easier to work with theoretically.
Dual Graphs
For a connected planar graph G with a fixed embedding, the dual graph G* is:
- One vertex in G* for each face of G.
- An edge in G* for each edge in G, connecting the two faces that the edge separates.
Properties:
- (G*)* = G (for connected planar graphs).
- |V(G*)| = f(G), |E(G*)| = |E(G)|, |faces(G*)| = |V(G)|.
- The dual of a tree is a single vertex with loops.
- Euler's formula for G and G* are consistent.
Applications:
- A spanning tree of G corresponds to a complementary spanning tree of G* (edges in MST of G ↔ edges not in co-tree of G*).
- The dual of a planar graph is used in shortest path algorithms on planar graphs.
- Map coloring is equivalent to vertex coloring the dual graph.
Graph Minor Theory
The Robertson-Seymour theorem (Graph Minor Theorem, proved over 20 papers from 1983-2004) states:
Every minor-closed family of graphs can be characterized by a finite set of forbidden minors.
A family F is minor-closed if: G ∈ F and H is a minor of G implies H ∈ F.
Examples:
- Planar graphs: Forbidden minors = {K₅, K₃,₃} (Wagner's theorem)
- Forests: Forbidden minor = {K₃} (a cycle)
- Outerplanar graphs: Forbidden minors = {K₄, K₂,₃}
- Graphs of treewidth ≤ k: Finite set of forbidden minors (but the set grows enormously)
Consequences: For any minor-closed property, there is a polynomial-time algorithm to test it (via forbidden minor testing). However, the algorithm may not be constructive — we may not know the forbidden minors explicitly.
Outerplanar Graphs
A graph is outerplanar if it can be drawn in the plane with all vertices on the boundary of the outer face.
Equivalent: Planar with no K₄ or K₂,₃ minor.
Properties:
- Outerplanar ⟹ χ(G) ≤ 3
- Outerplanar ⟹ treewidth ≤ 2
- Every outerplanar graph has a vertex of degree ≤ 2
Genus and Beyond
The genus of a graph G is the minimum number of handles that must be added to a sphere to embed G without crossings.
- Genus 0: Planar graphs (embeddable on sphere)
- Genus 1: Embeddable on torus (e.g., K₇)
- Higher genus: More complex surfaces
Euler's formula for genus g:
v - e + f = 2 - 2g
Heawood Conjecture (proved except for Klein bottle): The maximum chromatic number of graphs embeddable on a surface of genus g ≥ 1 is:
⌊(7 + √(1 + 48g)) / 2⌋
Crossing Number
The crossing number cr(G) is the minimum number of edge crossings in any drawing of G.
- cr(G) = 0 iff G is planar.
- cr(K₅) = 1, cr(K₆) = 3.
- Computing cr(G) is NP-hard.
Crossing Lemma (Ajtai-Chvátal-Newborn-Szemerédi, 1982): If e ≥ 4v, then:
cr(G) ≥ e³ / (64v²)
This has numerous applications in combinatorial geometry.
Planarity Testing Algorithms
Boyer-Myrvold Algorithm
O(V + E) time. Based on edge addition and embedding maintenance.
Left-Right Planarity Test
Also O(V + E). Uses a characterization based on DFS and T-edges vs B-edges.
Practical Approach
For implementation, Boyer-Myrvold is well-suited. Many graph libraries include planarity testing (e.g., Boost Graph Library, NetworkX, OGDF).
Real-World Applications
- Circuit board design (PCB): Minimizing wire crossings. Planar circuits require fewer layers.
- VLSI layout: Planar graphs model single-layer circuit designs. Non-planar requires via holes between layers.
- Map coloring: As discussed, planarity guarantees 4-colorability.
- Network visualization: Planar layouts are easier to read. Graph drawing algorithms often try to minimize crossings.
- Road networks: Road maps are approximately planar (bridges and tunnels are exceptions). This enables efficient algorithms: shortest paths in planar graphs can be computed in O(n) with preprocessing.
- Mesh generation: Finite element methods use planar (triangulated) meshes for simulation.
- Geographic information systems: Maps, boundaries, and spatial data are naturally planar.