Utility Graph — Definition, Formula & Examples
The utility graph is the complete bipartite graph , formed by connecting each of three "utility" vertices to each of three "house" vertices, giving 9 edges total. It is famous as the smallest example showing that not every graph can be drawn in a plane without edge crossings.
The utility graph is the complete bipartite graph whose vertex set is partitioned into two disjoint sets and , with an edge joining every vertex in to every vertex in . By Kuratowski's theorem, is one of two fundamental nonplanar graphs (the other being ).
How It Works
The classic puzzle asks: can you connect three houses to three utilities (water, gas, electricity) so that no pipes cross? You draw 6 vertices arranged in two rows of 3 and try to connect every top vertex to every bottom vertex without any edges intersecting. No matter how you rearrange the drawing, at least two edges must cross. This impossibility is guaranteed by the fact that is nonplanar, which can be verified using Euler's formula for planar graphs.
Worked Example
Problem: Use Euler's formula to prove the utility graph K₃,₃ cannot be planar.
Step 1: Count vertices and edges. has vertices and edges.
Step 2: For a connected planar graph, Euler's formula gives , so the number of faces would be .
Step 3: Since is bipartite, it has no odd cycles, so every face is bounded by at least 4 edges. Each edge borders at most 2 faces, so . Check: but . Since , we have a contradiction.
Answer: cannot be drawn in the plane without edge crossings; it is nonplanar.
Why It Matters
Kuratowski's theorem states that a graph is nonplanar if and only if it contains a subdivision of or . This makes the utility graph a cornerstone test case in circuit board design, network layout, and any problem where crossing-free embedding matters.
Common Mistakes
Mistake: Believing that because one particular drawing of K₃,₃ has crossings, you just haven't tried hard enough to find a crossing-free drawing.
Correction: Nonplanarity is a proven mathematical property of the graph itself, not a limitation of any single drawing. No rearrangement of vertices will eliminate all crossings.
