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.
| Assignment problem | |
|---|---|
| Name | Assignment problem |
| Developer | Kuhn, Hungarian algorithm |
| Introduced | 1930s–1950s |
| Input | Bipartite cost matrix |
| Output | Minimum-cost perfect matching |
| Complexity | Polynomial time |
Assignment problem
The assignment problem is a fundamental combinatorial optimization task that seeks a minimum-cost matching between two equal-sized sets, typically modeled as agents and tasks. Originating in early 20th-century operations research and graph theory, it links developments from Kuhn and Hungarian algorithm to later work in Hungary and United States research centers. It played a role in applications encountered by institutions such as Bell Labs, RAND Corporation, IBM, and in algorithmic theory advanced at MIT, Stanford University, and Princeton University.
The canonical formulation considers an n×n cost matrix C with rows indexed by members of one set and columns by members of another set; the objective is to choose n entries with distinct rows and columns minimizing total cost. Equivalent formulations appear as a weighted perfect matching problem on a complete bipartite graph K_{n,n} and as a linear program over permutation matrices with integrality guaranteed by total unimodularity. Historical expositions reference the work of Kuhn, Dantzig, Hungarian algorithm, and early input from researchers affiliated with National Bureau of Standards and Courant Institute.
Mathematical properties include total unimodularity of the constraint matrix, duality in linear programming with complementary slackness conditions, and connections to Birkhoff–von Neumann theorem on doubly stochastic matrices. Variants comprise the rectangular assignment (m×n) with unbalanced sizes, the assignment with forbidden pairs, the assignment with capacities, and the bottleneck assignment focusing on minimax criteria. Specialized models include the quadratic assignment problem linked to facility location and the three-index assignment connected to multiway matching; influential contributors to these areas include researchers from INFORMS, SIAM, and scholars such as Koopmans and Beckmann.
Exact polynomial-time algorithms include the Hungarian algorithm, successive shortest augmenting path methods, and primal-dual techniques grounded in linear programming and network flow theory. Implementations often rely on shortest path subroutines related to Dijkstra and potentials inspired by Bellman concepts. For dense instances, variants of the Hungarian method and cost-scaling algorithms developed in research groups at Bell Labs and AT&T yield practical performance; for sparse graphs, augmenting path methods and push–relabel adaptations are common. For integer-programming formulations, branch-and-bound and cutting-plane approaches were advanced by researchers at IBM and ETH Zurich.
The assignment problem is solvable in polynomial time; classical Hungarian algorithm runs in O(n^3), with improvements achieving O(n^2 log n + n m) and specialized implementations reaching near O(n^2) in practice for particular distributions. In contrast, generalizations like the quadratic assignment problem and the three-index assignment are NP-hard and have exponential worst-case complexity; hardness results were established in complexity theory circles at institutions such as Bell Labs and universities including UC Berkeley and Carnegie Mellon University. Average-case analyses and smoothed complexity studies have been conducted by teams at Microsoft Research and Google.
Applications span scheduling and personnel allocation in organizations like United Nations peacekeeping deployment and NATO logistics, transportation pairing used by UPS and FedEx, image feature matching in projects at NASA and European Space Agency, and resource allocation problems studied at World Bank and OECD. Other applications include assignment-based formulations in experimental design linked to the work of Fisher and supply chain optimizations utilized by Walmart and Amazon. In computational biology, matching formulations appear in sequence assembly and protein docking projects at Sanger Institute and Broad Institute.
A simple 3×3 example: given cost matrix C = 8, 4, 7], [6, 5, 9], [3, 8, 2, the optimal assignment selects entries that minimize sum; algorithmic traces of the Hungarian method show successive row and column reductions, construction of a zero-cover, and augmentation steps familiar from courses at Harvard, Yale, and Columbia University. Numerical libraries implementing these procedures include packages from Netlib, GNU Project, and commercial solvers from Gurobi and CPLEX which provide benchmarks on datasets used in competitions organized by INFORMS.
Extensions encompass the quadratic assignment problem (QAP), multi-index assignment, stochastic assignment under uncertainty studied at Stanford University and Harvard, and online assignment encountered in ride-hailing platforms developed by Uber and Lyft. Other generalizations integrate capacity constraints leading to transportation problems linked historically to Kantorovich and continuous relaxations used in optimal transport theory advanced by groups at NYU and EPFL. Decompositions and approximation schemes have been advanced by teams at ETH Zurich, University of Waterloo, and INRIA.