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.
| Minimum-cost flow problem | |
|---|---|
| Name | Minimum-cost flow problem |
| Input | Directed graph, capacities, costs, supplies/demands |
| Output | Flow with minimum total cost meeting supplies/demands |
| Complexity | Varies by algorithm; polynomial-time solvable |
Minimum-cost flow problem The minimum-cost flow problem is a network optimization task that seeks a flow that satisfies node supplies and demands while minimizing total edge cost. It generalizes classical network problems and connects to models studied in operations research, combinatorics, and theoretical computer science. The problem has rich interactions with influential figures and institutions in optimization and algorithm design.
In a capacitated directed graph with nodes and directed edges, each node may act as a supplier or a demander and each edge has an associated nonnegative cost per unit and a capacity limit. A feasible flow conserves net supply at each node and respects edge capacities; among feasible flows one seeks one of minimum total cost. This setting extends frameworks explored by George Dantzig, John von Neumann, Leonid Kantorovich, Richard Karp, and formalized in foundational work from Bell Labs and RAND Corporation on network flows and linear programming.
The problem is commonly cast as a linear program: minimize the sum over edges of cost times flow subject to flow conservation constraints at each node and capacity bounds on each edge. Dual variables correspond to node potentials, connecting optimality conditions to complementary slackness and to shortest-path structures underlying optimal augmentations. This LP perspective ties the problem to results in Linear programming, George Dantzig's simplex method, and polyhedral studies by Václav Chvátal and Martin Grötschel.
Polynomial-time solvability arises via reductions to minimum-cost circulation and network simplex, successive shortest path, cycle canceling, and cost-scaling methods. The Network simplex algorithm adapts the simplex method to spanning-tree bases, while the Successive shortest path algorithm repeatedly augments along shortest paths in a residual network using potentials from Dijkstra or Bellman–Ford techniques. Cost-scaling algorithms, influenced by work of Andrew Goldberg and Robert Tarjan, achieve strongly polynomial or near-linear performance in practice. Complexity analyses reference seminal contributions by Jack Edmonds, Richard Karp, and later developments by Éva Tardos on combinatorial optimization.
Several canonical problems are special cases: the maximum flow and minimum cut problems, the assignment problem and bipartite matching, and the transportation problem studied by Fisher-type models. Variants include the transshipment problem, capacitated transshipment, min-cost circulation, multi-commodity flow motivated by Claude Shannon's information networks, and stochastic or dynamic minimum-cost flows influenced by models from W. Edwards Deming-style supply chain literature. Integer requirements yield the integer minimum-cost flow, connected to integrality theorems by T. C. Hu and total unimodularity results attributed to Hassler Whitney-related matroid theory.
Practical applications span logistics and routing in industries such as freight and airline scheduling, production planning at firms highlighted by Toyota and Ford Motor Company, and telecommunication capacity planning in networks inspired by work at AT&T and Bell Labs. In computer science and operations research, it underpins resource allocation in cloud computing by companies like Amazon Web Services and Google and revenue management in service industries studied by scholars at MIT and Stanford University. Other applications include energy distribution in grids influenced by projects at Siemens and General Electric, and transportation planning in metropolitan agencies such as the Metropolitan Transportation Authority.
Classic textbook instances illustrate small graphs where cheapest augmenting paths yield optimal flows; textbooks by Alexander Schrijver and Ahuja, Magnanti, and Orlin present canonical examples with step-by-step augmentations. Benchmark instances used in experimental comparisons originate from datasets curated by research groups at Cornell University and University of California, Berkeley. Visual illustrations often depict residual networks, negative-cost cycles discovered in cycle-canceling methods, and potential-labeled shortest-path trees employed in successive shortest path demonstrations.
Category:Network flow problems