Randomized rounding: a technique for provably good algorithms and algorithmic proofs
Combinatorica - Theory of Computing
Probabilistic checking of proofs: a new characterization of NP
Journal of the ACM (JACM)
An O(log k) Approximate Min-Cut Max-Flow Theorem and Approximation Algorithm
SIAM Journal on Computing
Proof verification and the hardness of approximation problems
Journal of the ACM (JACM)
SIAM Journal on Computing
Random Sampling in Cut, Flow, and Network Design Problems
Mathematics of Operations Research
Multicommodity max-flow min-cut theorems and their use in designing approximation algorithms
Journal of the ACM (JACM)
Approximate Max-Flow Min-(Multi)Cut Theorems and Their Applications
SIAM Journal on Computing
Minimizing Congestion in General Networks
FOCS '02 Proceedings of the 43rd Symposium on Foundations of Computer Science
ICALP '93 Proceedings of the 20th International Colloquium on Automata, Languages and Programming
The all-or-nothing multicommodity flow problem
STOC '04 Proceedings of the thirty-sixth annual ACM symposium on Theory of computing
Multicommodity flow, well-linked terminals, and routing problems
Proceedings of the thirty-seventh annual ACM symposium on Theory of computing
Hardness of the undirected edge-disjoint paths problem
Proceedings of the thirty-seventh annual ACM symposium on Theory of computing
New hardness results for congestion minimization and machine scheduling
Journal of the ACM (JACM)
Hardness of the Undirected Congestion Minimization Problem
SIAM Journal on Computing
The geometry of graphs and some of its algorithmic applications
SFCS '94 Proceedings of the 35th Annual Symposium on Foundations of Computer Science
Expander flows, geometric embeddings and graph partitioning
Journal of the ACM (JACM)
Edge Disjoint Paths in Moderately Connected Graphs
SIAM Journal on Computing
Approximation Algorithms for the Edge-Disjoint Paths Problem via Raecke Decompositions
FOCS '10 Proceedings of the 2010 IEEE 51st Annual Symposium on Foundations of Computer Science
Routing in undirected graphs with constant congestion
STOC '12 Proceedings of the forty-fourth annual ACM symposium on Theory of computing
Reconstructing edge-disjoint paths
Operations Research Letters
Hi-index | 0.00 |
We study an integral counterpart of the classical Maximum Concurrent Flow problem, that we call Integral Concurrent Flow (ICF). In the basic version of this problem (basic-ICF), we are given an undirected n-vertex graph $G$ with edge capacities c(e), a subset T of vertices called terminals, and a demand D(t,t') for every pair (t,t') of the terminals. The goal is to find a maximum value λ, and a collection P of paths, such that every pair (t,t') of terminals is connected by ⌊ λ ⋅ D(t,t')⌋ paths in P, and the number of paths containing any edge e is at most c(e). We show an algorithm that achieves a poly log n-approximation for basic-ICF, while violating the edge capacities by only a constant factor. We complement this result by proving that no efficient algorithm can achieve a factor α-approximation with congestion c for any values α,c satisfying α ⋅ c=O(log log n/log log log n), unless NP ⊆ ZPTIME(npoly log n). We then turn to study the more general group version of the problem (group=ICF), in which we are given a collection (S1,T1),...,(Sk,Tk)} of pairs of vertex subsets, and for each 1 ≤ i ≤ k, a demand Di is specified. The goal is to find a maximum value λ and a collection P of paths, such that for each i, at least ⌊ λ ⋅ Di⌋ paths connect the vertices of Si to the vertices of Ti, while respecting the edge capacities. We show that for any 1 ≤ c ≤ O(log log n), no efficient algorithm can achieve a factor O(n1/(22c+3))-approximation with congestion c for the problem, unless NP ⊆ DTIME(nO(log log n)). On the other hand, we show an efficient randomized algorithm that finds a poly log n-approximate solution with a constant congestion, if we are guaranteed that the optimal solution contains at least D ≥ k poly log n paths connecting every pair (Si,Ti).