The Straight-Line RAC Drawing Problem is NP-Hard

Recent cognitive experiments have shown that the negative impact of an edge crossing on the human understanding of a graph drawing, tends to be eliminated in the case where the crossing angles are greater than 70 degrees. This motivated the study of RAC drawings, in which every pair of crossing edges intersects at right angle. In this work, we demonstrate a class of graphs with unique RAC combinatorial embedding and we employ members of this class in order to show that it is NP-hard to decide whether a graph admits a straight-line RAC drawing.

💡 Research Summary

The paper investigates the computational complexity of producing straight‑line RAC (Right‑Angle Crossing) drawings, a class of graph visualizations in which every pair of crossing edges meets at a right angle. The motivation stems from cognitive experiments showing that the detrimental effect of edge crossings on human comprehension largely disappears when the crossing angle exceeds about 70°, suggesting that right‑angle crossings are visually benign. While previous work has examined RAC drawings with curved edges or shown that certain restricted graph families admit straight‑line RAC drawings in polynomial time, the complexity for general graphs remained open.

The authors introduce a novel family of graphs that possess a unique RAC combinatorial embedding: any embedding that respects the RAC condition is forced to be exactly the same up to planar isotopy, and any deviation necessarily violates the right‑angle requirement. This uniqueness is achieved by carefully arranging triangles, quadrilaterals, and forced crossing gadgets so that each crossing must occur at a prescribed location and orientation.

Using these gadgets, the paper constructs a polynomial‑time reduction from Planar 3‑SAT, a known NP‑complete problem, to the straight‑line RAC drawing decision problem. Each Boolean variable is represented by a “variable gadget” that can be drawn in exactly two ways, corresponding to the true/false assignment. Each clause is represented by a “clause gadget” that connects to three variable gadgets; the clause gadget can be embedded in a RAC‑compliant way if and only if at least one of its incident variable gadgets is in the “true” configuration. All connecting edges are straight lines, and the design guarantees that every crossing forced by the construction occurs at a right angle, thanks to the unique embedding property.

The reduction proves two crucial points. First, if the original Planar 3‑SAT instance is satisfiable, the assembled graph admits a straight‑line RAC drawing by selecting the appropriate variable configurations. Second, if the assembled graph has a straight‑line RAC drawing, the embedding of each gadget reveals a satisfying assignment for the 3‑SAT instance. Consequently, deciding whether an arbitrary graph admits a straight‑line RAC drawing is NP‑hard. The authors also argue that the problem lies in NP, because a candidate drawing can be verified in polynomial time by checking planarity, straight‑line edges, and right‑angle crossings, establishing NP‑completeness.

Beyond the hardness result, the paper discusses practical implications for graph drawing software. Since optimal straight‑line RAC layouts cannot be computed efficiently for general graphs, designers must rely on heuristics, approximation schemes, or restrict themselves to graph families known to be tractable. The introduction of the unique RAC embedding concept opens new avenues for analyzing other angle‑constrained drawing problems, such as those involving limited slope sets or fixed angular resolution.

In conclusion, the work settles a long‑standing open question by proving that the straight‑line RAC drawing problem is NP‑hard (indeed NP‑complete), thereby delineating the theoretical limits of aesthetically pleasing, cognitively friendly graph visualizations. Future research directions include identifying broader tractable subclasses, developing effective heuristic algorithms, and empirically testing whether right‑angle crossings continue to aid human comprehension in large‑scale, automatically generated drawings.