Hamiltonian path problem

Hamiltonian path problem The Hamiltonian path problem asks whether a given finite graph contains a simple path that visits every vertex exactly once, and it is a central decision problem in theory of computation and combinatorics. The problem is named after William Rowan Hamilton and is closely related to the Travelling Salesman Problem and the concept of Hamiltonian cycle, with deep connections to Leonhard Euler's earlier work on bridges problems and later developments in graph algorithms and computational complexity theory.

Definition and basic concepts

A Hamiltonian path in a finite graph is a simple path that contains every vertex exactly once; when the path is closed it is a Hamiltonian cycle or Hamiltonian circuit, a notion studied by William Rowan Hamilton in the context of the Icosian Game. Formal definitions use vertices and edges from graph instances, distinguishing between directed and undirected cases and between weighted and unweighted variants. Closely related concepts include Eulerian trail from Leonhard Euler's problems, path cover in flow networks, and graph connectivity measures such as vertex connectivity and edge connectivity. Classic sufficient conditions for Hamiltonicity include theorems by Gabriel Andrew Dirac and Bondy and Ore, linking degree constraints to the existence of Hamiltonian cycles.

Computational complexity and NP-completeness

The Hamiltonian path problem is one of Karp's original 21 NP-complete problems from 1972 and is a canonical NP-complete decision problem in computational complexity theory. The decision version—does a graph contain a Hamiltonian path?—is in NP and, because of polynomial-time reductions from problems such as SAT and 3-SAT, is NP-hard; therefore it is NP-complete unless P=NP. The optimization variant, asking for a longest simple path, is NP-hard and often treated in the context of approximation algorithms and parameterized complexity such as fixed-parameter tractability results by researchers in parameterized complexity theory. Hardness results connect to the Cook–Levin framework and reductions involving classical problems studied at institutions like Bell Labs and IBM Research.

Algorithms and heuristics

Exact algorithms include exhaustive search and backtracking refined by pruning strategies used in implementations at Bell Labs and in academic work by scholars from MIT and Stanford University. Dynamic programming over subsets, notably the Held–Karp algorithm developed in part at RAND Corporation contexts, solves the Travelling Salesman Problem and hence Hamiltonian path in O(n^2 2^n) time. Branch-and-bound and cutting-plane methods were advanced by researchers at IBM Research and AT&T for practical instances. Heuristics and metaheuristics—such as simulated annealing developed in Bell Labs contexts, genetic algorithms popularized in work at Johns Hopkins University and University of Michigan, and ant colony optimization pioneered by teams including researchers at IRIDIA—provide practical approximations. Specialized algorithms exploit structure revealed by research from Carnegie Mellon University and École Normale Supérieure; implementations often integrate with libraries originating at INRIA and Los Alamos National Laboratory.

Special graph classes and variations

Polynomial-time solvable cases include Hamiltonicity tests on trees (trivial), outerplanar graphs, threshold graphs, and graphs of bounded treewidth where Courcelle's theorem and dynamic programming techniques apply; such results originate in work at University of Oxford and ETH Zurich. Planar graphs and bipartite graphs have specialized criteria and algorithms studied by researchers at Princeton University and University of Cambridge. Variations include the directed Hamiltonian path problem analyzed in contexts at Caltech, Hamiltonian decomposition studied by researchers connected to Cayley graph theory, and the colored or constrained Hamiltonian path problems appearing in work at University of Toronto and Tsinghua University. Parameterized variants consider parameters like path length or feedback vertex set size and have been developed in the parameterized complexity community including groups at University of Bergen and MPI-SWS.

Applications and practical uses

Applications span classical industrial and scientific problems: routing and scheduling in transportation systems studied at MIT and Georgia Institute of Technology, genome assembly and sequence reconstruction in bioinformatics groups at Broad Institute and Sanger Institute, circuit layout and testing in semiconductor industry labs such as Intel and TSMC, and puzzle design and analysis exemplified by works connected to MAA activities and Recreational mathematics research at University of Cambridge. Variants underpin aspects of cryptography research at RSA Laboratories and NIST for hardness assumptions, and they appear in planning problems in robotics developed at Carnegie Mellon University and ETH Zurich. Practical solvers are embedded in software suites from IBM and academic toolchains at University of Illinois Urbana–Champaign used for benchmarking and comparative studies.

Category:Graph theory problems