Summary.
The maximum flow problem is a fundamental problem in graph theory and combinatorial optimization with a variety of important applications. Known distributed algorithms for this problem do not tolerate faults or adjust to dynamic changes in network topology. This paper presents a distributed self-stabilizing algorithm for the maximum flow problem. Starting from an arbitrary state, the algorithm computes the maximum flow in an acyclic network in finitely many steps. Since the algorithm is self-stabilizing, it is inherently tolerant to transient faults. It can automatically adjust to topology changes and to changes in other parameters of the problem. The paper presents results obtained by extensively experimenting with the algorithm. Two main observations based on these results are (1) the algorithm requires fewer than n 2 moves for almost all test cases and (2) the algorithm consistently performs at least as well as a distributed implementation of the well-known Goldberg-Tarjan algorithm for almost all test cases. The paper ends with the conjecture that the algorithm correctly computes a maximum flow even in networks that contain cycles.
Similar content being viewed by others
Author information
Authors and Affiliations
Additional information
Received: October 1995 / Accepted: February 1997
Rights and permissions
About this article
Cite this article
Ghosh, S., Gupta, A. & Pemmaraju, S. A self-stabilizing algorithm for the maximum flow problem. Distrib Comput 10, 167–180 (1997). https://doi.org/10.1007/s004460050034
Issue Date:
DOI: https://doi.org/10.1007/s004460050034