This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| Graph isomorphism problem | |
|---|---|
| Name | Graph isomorphism problem |
| Field | Mathematics, Theoretical computer science |
| Introduced | 1941 |
Graph isomorphism problem
The Graph isomorphism problem asks whether two finite graphs are structurally identical, i.e., whether there exists a bijection between their vertex sets preserving adjacency. This decision problem has motivated research across Évariste Galois, Kurt Gödel, Alan Turing, John von Neumann, Paul Erdős and influenced work at institutions such as Institute for Advanced Study, Bell Labs, MIT, Princeton University, Harvard University and Stanford University. It connects to results in complexity theory from Stephen Cook, Richard Karp, Leslie Valiant, Andrew Yao and intersects algorithmic advances from László Babai, Joan Feigenbaum, Michael Sipser, Richard J. Lipton and Neil Immerman.
Formally, given two undirected graphs G and H, the problem asks whether there exists a bijection f: V(G) → V(H) such that {u,v} ∈ E(G) iff {f(u),f(v)} ∈ E(H). The notion of isomorphism is central in combinatorics studied by William Tutte, Claude Berge, Pierre Deligne and in algebraic graph theory influenced by Emil Artin, Niels Henrik Abel, Arthur Cayley and Issai Schur. Variants consider directed graphs, colored vertices, labeled edges, and relational structures analyzed by Alfred Tarski, S. C. Kleene and Alonzo Church. Early formalizations appear in work by Frank Harary and in enumerative combinatorics advanced by George Pólya.
The complexity status of the problem resists classic classification: it is in NP but not known to be NP-complete under standard assumptions, a position highlighted in surveys by Richard Lipton and debated at conferences such as STOC and FOCS. Results include quasi-polynomial time algorithms by László Babai and group-theoretic techniques inspired by Élie Cartan and Camille Jordan. Connections to group theory trace to Galois theory and computational group algorithms developed by Charles Sims, Ákos Seress and John Dixon. Lower bounds and hardness results relate to works by Mihalis Yannakakis, Leonid Levin and Stephen Cook. The problem lies in complexity classes studied by Peter Shor, Scott Aaronson, Thomas Hales and links to descriptive complexity from Neil Immerman and Anuj Dawar.
Classical algorithms include backtracking and refinement methods such as the Weisfeiler–Leman algorithm developed by Boris Weisfeiler and Andrei Leman and refined in literature by László Babai and Martin Grohe. Group-theoretic approaches leverage permutation group algorithms from Évariste Galois lineage, with practical implementations using algorithms by Ákos Seress and theoretical foundations by Charles Sims and Dixon. Spectral methods exploit eigenvalues and eigenvectors building on work by Issai Schur, Alfred North Whitehead and Stefan Banach; algebraic invariants trace to Emmy Noether and Hermann Weyl. Refinement and partitioning heuristics were advanced by Donald Knuth and coded in toolkits from Bell Labs and IBM Research. Recent progress integrates machine learning ideas influenced by Yann LeCun and Geoffrey Hinton with combinatorial optimization traditions from Jack Edmonds.
Polynomial-time solvable classes include trees (classical results by Kruskal and Robert Tarjan), planar graphs (studied by Klaus Wagner, Kazimierz Kuratowski, William Tutte), bounded degree graphs (results by Eugene Luks) and graphs of bounded genus (related to Heinrich Heesch). Interval graphs and chordal graphs tie to work by Martin Golumbic, while cographs and permutation graphs feature in research by D. G. Corneil and Graham Pride. Strong structural theorems for graph minors come from Paul Seymour and Neil Robertson, impacting isomorphism algorithms for excluded-minor families. Random graph models studied by Paul Erdős and Alfréd Rényi yield almost-everywhere easy instances, linked to probabilistic combinatorics from Béla Bollobás.
Software packages implementing isomorphism routines include systems from Brendan McKay (nauty), development groups at University of Waterloo, Rensselaer Polytechnic Institute, INRIA and commercial efforts at IBM Research. Applications appear in cheminformatics for molecular pattern matching used by Merck and Pfizer, in electronic design automation in companies like Intel and Cadence Design Systems, and in database deduplication in firms such as Oracle Corporation and Microsoft Corporation. Structural comparison tasks arise in bioinformatics with tools developed by labs at National Institutes of Health and European Bioinformatics Institute; pattern recognition applications intersect with work by Herbert A. Simon and Allen Newell.
Closely related decision and search problems include subgraph isomorphism with NP-complete stature proved in reductions by Richard Karp and Stephen Cook, graph automorphism with group-theoretic ties explored by László Babai and Joan Feigenbaum, and the canonical labeling problem central to database indexing researched by Donald Knuth and Brendan McKay. Connections extend to isomorphism of algebraic structures studied by Emmy Noether and Richard Dedekind, and to constraint satisfaction problems linked to work by Avi Wigderson and Mihalis Yannakakis. Complexity-theoretic frameworks involving interactive proofs and zero-knowledge proofs bring in contributions from Oded Goldreich, Shafi Goldwasser and Silvio Micali.