A. V. Goldberg and R. E. Tarjan, "A new approach to the maximum flow problem," J. ACM, pp. 921--940, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....T) c(u, v) The minimum s t cut problem is to find a cut C with the smallest cost. Due to the theorem of Ford and Fulkerson [13] this is equivalent to computing the maximum flow from the source to sink. There are many algorithms which solve this problem in polynomial time with small constants [1,7,16]. It is convenient to denote a cut C = S, T by a labeling f mapping from the set of the vertices t to 1 where f(v) 0meansthatv S,andf(v) 1meansthatv T . We will use this notation later. Note that a cut is a binary partition of a graph; viewed as a labeling, it is a binary valued ....

A. Goldberg and R. Tarjan. A new approach to the maximum flow problem. Journal of the Association for Computing Machinery, 35(4):921--940, October 1988.

....[1990] gave an O( A log( V A ) algorithm for finding a blocking flow in an acyclic directed graph, implying an O( V A ) algorithm for finding a maximum flow in any directed graph. An alternative approach finding a maximum flow in time O( V A ) was described in Goldberg and Tarjan [1988]. For surveys on maximum flow algorithms, see Goldberg, Tardos, and Tarjan [1990] and Ahuja, Magnanti, and Orlin [1993] This can be done recursively as follows (cf. Knuth [1968] Tarjan [1974] If # (s) then the ordering is trivial. If # (s) choose (s, v) s) and order the ....

....can show that the running time of this algorithm is O(M (m n log n) where M is the value of a maximum flow (assuming all capacities to be integer) So it is not polynomial time. At the moment of writing, the asymptotically fastest method for finding a minimum cost maximum flow was designed by Orlin [1988,1993] and runs in O(m log n(m n log n) time. In a similar way one can describe a minimum cost circulation algorithm. For more about network flows we refer to the books of Ford and Fulkerson [1962] and Ahuja, Magnanti, and Orlin [1993] Application 4.3: Minimum cost transportation problem. ....

A.V. Goldberg, R.E. Tarjan, A new approach to the maximum-flow problem, Journal of the Association for Computing Machinery 35 (1988) 921--940.

....c(u, v) The minimum s t cut problem is to find a cut C with the smallest cost. Due to the theorem of Ford and Fulkerson [9] this is equivalent to computing the maximum flow from the source to sink. There are many algorithms which solve this problem in polynomial time with small constants [1, 11]. It is convenient to denote a cut C = S, T by a labeling f mapping from the set of the nodes V s, t to 1 where f(v) 0meansthatv S,and f(v) 1meansthatv T . We will use this notation later. 3 Defining graph representability Let us consider a graph with terminals s and t,thusV ....

A. Goldberg and R. Tarjan. A new approach to the maximum flow problem. Journal of the Association for Computing Machinery, 35(4):921--940, October 1988.

.... formulated dual problem. Section 2 of this paper gives an overview and partial history of these methods. Section 3 examines in detail what is perhaps the generic algorithm of the class, the e relaxation method [7] Section 4 develops some basic serial complexity analysis tools for this algorithm [28, 29, 8], also addressing the special case of maximum flow problems. Section 5 combines this analysis with the notion of scaling [30 32, 19, 24, 41, 5] yielding a polynomial (O(N 3 log NC) serial algorithm for the minimum cost flow problem (N is the number of nodes, and C the largest absolute value of ....

....as those depicted in Fig. 5. If e d N, the algorithm will still terminate with a feasible flow, but this flow may not be optimal. The Goldberg Tarjan maximum flow method Another important algorithm belonging to the dual coordinate step class is the maximum flow method of Goldberg and Tarjan [28, 29]. This algorithm was developed roughly concurrently with, and entirely independently from, the auction algorithm and the RELAX family of codes. The original motivation for this algorithm seems to have been quite different from the theory we emphasize in this paper; it appears to have been ....

[Article contains additional citation context not shown here]

A.V. Goldberg and R.E. Tarjan, "A new approach to the maximum flow problem," Proc. 18th ACM STOC, 1986, pp. 136-146.

.... Gallo, Grigoriadis, and Tarjan [3] It is obtained by reducing the densest subgraph problem to a parametric min cut problem and then solving it using a parametric max flow algorithm whose running time is the same as the running time of the non parametric max flow algorithm of Goldberg and Tarjan [5]. Of more practical interest is a much simpler linear time 2 approximation algorithm for the densest subgraph problem which is a slight modification of an algorithm mentioned by Kortsarz and Peleg [7] This algorithm iteratively removes a vertex of minimum degree from the graph. This generates a ....

A. V. Goldberg and R.E. Tarjan. A new approach to the maximum flow problem. J. Assoc. Comput. Mach., 35:921--940, 1988.

.... is done by solving a maximum flow problem on an auxiliary undirected graph with capacities depending on the current solution (x # , y # ) and on k (see Magnanti and Wolsey [15] for this construction for the MSTP) The max flow algorithm used in our implementation is taken from Goldberg and Tarjan [11]. For k K the auxiliary graph G( V , E) is defined as follows. Let V = V # t and E = E # v : v # v, t : v . The capacity for every edge e E # is x # e 2, the capacity of an edge E is equal to x # (#(v) 2 and the capacity of an edge t E is ....

A.V. Goldberg, R.E. Tarjan. A New Approach to the Maximum Flow Problem. Journal of Association for Computing Machinery 35 (1988) 921--940.

.... Gallo, Grigoriadis, and Tarjan [3] It is obtained by reducing the densest subgraph problem to a parametric min cut problem and then solving it using a parametric max flow algorithm whose running time is the same as the running time of the non parametric max flow algorithm of Goldberg and Tarjan [5]. Of more practical interest is a much simpler linear time 2 approximation algorithm for the densest subgraph problem which is a slight modification of an algorithm mentioned by Kortsarz and Peleg [7] This algorithm 6 iteratively removes a vertex of minimum degree from the graph. This generates ....

A. V. Goldberg and R.E. Tarjan. A new approach to the maximum flow problem. J. Assoc. Comput. Mach., 35:921--940, 1988.

....optimal solution to Problem EER and that the complexity of the Binary Iterative Algorithm is logarithmic in the network life time. Since there is a single destination node, an instance of Problem CMF can be solved by using binary search with a max flow algorithm (e.g. the preflow push algorithm [13])J Problem CIVIF can also be solved by using binal search with a version of the approximate algorithm presented in [3] Specifically, if for a given set of demands (i.e. for a given T) there exists a feasible flow (e.g. flow satisfying (8) 12) it can be found by a max flow algorithm. Thus, in ....

....T) At each iteration of the binary search (i.e. for a given T) a max flow algorithm is executed. It is executed in a network with a super origin node which is connected to every badge node i by a link whose capacity is tiT. Since the complexity of a max flow algorithm is O(n 3) 1, p. 240] [13], the number of steps required to find a solution to Problem CMF by Algorithm BMF is O(n 3 log Tmx ) 15) where Tmax is the maximal possible value of network lifetime (T) It can be shown that for a network of badges and the resuiting Problem CMF, the value of Tmax is bounded by n times the ....

[Article contains additional citation context not shown here]

A. V. Goldberg and R. E. Tarjan, "A New Approach to the Maximum Flow Problem", J. oftheACM, Vol. 35, pp. 921-940, Oct. 1988.

....Dijkstra s algorithm twice, the second execution is done on a much reduced network that often contains just a single path, increasing the execution time by far less than a factor of two. The WSUM MAX approximation computes several maxflows using Goldberg and Tarjan s preflow push algorithm [GT86] The average execution time is 3000s, which is more than a factor of 20 slower than the other algorithms. The initial computation of PBR s profile requires 30,000s and each individual path computation requires another 120s. Thus, the initial e#ort is amortized after about 10 path computations, ....

Andrew V. Goldberg and Robert E. Tarjan. A new approach to the maximum flow problem. In Proceedings of the Eighteenth Annual ACM Symposium on Theory of Computing, pages 136--146, 1986.

....Schrijver[16] minimizes f by calling the subroutine O (n 6) times. We will present another algorithm that calls it O(n ) times. A Push Relabel Algorithm for SFM We now describe the push relabel algorithm for SFM. The push relabel approach was introduced for network flows by Goldberg and Tarjan [10], and is among the most efficient known algorithms for maximum flow. It has been applied to polymatroid intersection, a problem equivalent to maximum submodular flow, by Fujishige and Zhang [9] The algorithm maintains x B(f) as a convex combina tion x = i )tiyi of extreme bases Yi and a directed ....

A. . Goldberg and R. E. Tarjan. A new approach to the maximum flow problem. Journal of the ACM, 35:921-940, 1988.

....on the problem of finding feasible flows for LPA in a loading configuration graph D = V; L; A) instead of minimum cost feasible flows. We assume the reader is familiar with the augmenting path algorithm by Ford and Fulkerson [FF62] and with the preflow push algorithm by Goldberg and Tarjan [GT88] Augmenting Paths. Let e (x) denote the total positive excess in the current flow x, i.e. x) i2V;k2K je ik (x)j. Consider the class of problems LPA for which d = 1 on a loading configuration graph D = V; L; A) This class is equal to network flow on D = V; A) if we ignore the ....

Goldberg, A. V. and Tarjan, R. E. A new approach to the maximum flow problem. J. Assoc. Comput. Mach. 35 (1988), 921--940. Also in Proc. 18th ACM Symp. on Theory of Comp., pages 136--146, 1986.

....to augmenting path algorithms of Tardos, Tovey, and Trick [18] We design a simpler push relabel framework for SFM that reduces the number of subroutine calls by a factor of n. The resulting algorithm runs in O(n ff n 8) time. The push relabel framework was introduced by Goldberg and Tarjan [9] for the maximum flow problem. Subsequently, it was applied to polymatroid intersection by Fujishige and Zhang [7] Gallo, Grigoriadis, and Tarjan [8] extended the push relabel algorithm to solve monotone para metric maximum flow problems with no increase in time complexity. Iwata, Murota, and ....

....of f. Otherwise, the set W of vertices reachable from N by the arcs in At is a minimizer of f. We call this variant Reverse Push Relabel. 3 Parametric Submodular Function Minimization Gallo, Grigoriadis, and Tarjan [8] modify the maximum flow push relabel algorithm of Goldberg and Tarjan [9] to solve a parametric network flow problem. They consider a flow network with arc capacities co that are functions of a parameter 0: For arc a leaving the source, co(a) is increasing in 0; for a entering the sink, co(a) is decreasing in 0; all other arcs have constant capacities. This is called a ....

A. V. Goldberg and R. E. Tarjan, A new approach to the maximum flow problem, J. A CM, 35 (1988), 921-940.

....when # # (x, y) is. The rest follows from Lemma 1. Lemma 3 Let # be as defined in Assumption F1. Given # as the input, we can find an x opt(#) in O(# log #) time. Proof: Given a digraph G = V, E) as input, the Goldberg Tarjan maximum flow algorithm takes O( V E log( V E ) time [14]. We first apply Lemma 2 to # to obtain G . We next use this maximum flow algorithm to find a minimum s t cut in G and then an optimal x from this cut. All these steps take O(# log #) total time. 3.3 A Compact Representation of Minimum Cuts A given # may have more than one fittest ....

A. V. Goldberg and R. E. Tarjan. A new approach to the maximum-flow problem. Journal of the ACM, 35(4):921--940, Oct. 1988.

....an s t maximum flow. In 1961, Gomory and Hu showed how to solve the minimum cut problem with n 1 s t minimum cut computations. Subsequently there was much progress in computing maximum flows, but no one has yet been able to prove a time bound better than O(nm) for any of the best algorithms [1, 9, 10, 25, 41]. Hence we cannot give a bound better than O(n m) for the Gomory Hu algorithm. Gomory Hu stood as the best algorithm for the problem until 1989, when Nagamochi and Ibaraki [47] showed how to find a minimum cut without using maximum flows. Their algorithm (which we will call NI) runs in O(n(m n ....

A. V. Goldberg and R. E. Tarjan. A New Approach to the Maximum Flow Problem. J. Assoc. Comput. Mach., 35:921--940, 1988.

....by a traversal of the implication graph. The time complexity of the traversal is O(V E) Here V and E are the number of vertices and edges of the implication graph respectively. The time complexity for finding a min cut of the implication graph can be implemented in O(VElg(V 2 E) time [17]. A good learning scheme should reduce the number of decisions needed to solve certain problems as much as possible. The effectiveness of a learning scheme is very hard to determine a priori. Generally speaking, a shorter clause Microprocessor Formal Verification[19] Bounded Model Checking [18] ....

A. Goldberg, R. Tarjan. "A new approach to the maximum flow problem," Proceedings of the eighteenth annual ACM Symposium on Theory of Computing, 1986.

.... O(k) In order to compare our experimental results with the theoretical bounds, we characterize our four test programs in terms of the parameters k and L, not counting the implicit fake r lock used by the detection algorithms: maxflow: A maximum flow code based on Goldberg s push relabel method [11]. Each vertex in the graph contains a lock. Parallel threads perform simple This chapter is based on jointwork published in [3] parameters time (sec. slowdown program input k L original All Sets Brelly All Sets Brelly maxflow sparse 1K 2 32 0.05 30 3 590 66 sparse 4K 2 64 0.2 484 14 2421 68 ....

Andrew V. Goldberg and Robert E. Tarjan. A new approach to the maximum flow problem. In Proceedings of the Eighteenth Annual ACM Symposium on Theory of Computing, pages 136--146, Berkeley, California, 28--30 May 1986.

....an s t maximum flow. In 1961, Gomory and Hu showed how to solve the minimum cut problem with n 1s tmini mum cut computations. Subsequently there was much progress in computing maximum flows, but no one has yet been able to prove a time bound better than O#nm# for any of the best algorithms [1, 9, 10, 25, 41]. Hence we cannot give a bound better than O#n m# for the Gomory Hu algorithm. Gomory Hu stood as the best algorithm for the problem until 1989, when Nagamochi and Ibaraki [47] showed how to find a minimum cut without using maximum flows. Their algorithm (which we will call NI) runs in O#n#m n ....

A. V. Goldberg and R. E. Tarjan. A New Approach to the Maximum Flow Problem. J. Assoc. Comput. Mach., 35:921--940, 1988.

....Dijkstra s algorithm twice, the second execution is done on a much reduced network that often contains just a single path, increasing the execution time by far less than a factor of two. The WSUM MAX approximation 26 computes several max flows using Goldberg and Tarjan s preflow push algorithm [21]. The average execution time is 3000s, which is more than a factor of 20 slower than the other algorithms. The initial computation of PBR s profile requires 30,000s and each individual path computation requires another 120s. Thus, the initial e#ort is amortized after about 10 path computations, ....

A. V. Goldberg, R. E. Tarjan, A new approach to the maximum flow problem, in: Proceedings of the Eighteenth Annual ACM Symposium on Theory of Computing, 1986, pp. 136--146. 30

....that jW j is at most nk. Ogier uses one maximum flow computation on a graph with at most nk vertices at each recursive step, yielding an O(n 4 k 4 ) algorithm to compute W . Substituting the fastest strongly polynomial maximum flow algorithm the preflowpush algorithm of Goldberg and Tarjan [5], Ogier s approach leads to an O(k 2 mn log(kn 2 =m) algorithm to compute x . 2 A Faster Algorithm In this section, we discuss the main contribution of this paper which is a generalization of the parametric maximum flow algorithm of Gallo, Grigoriadis, and Tarjan [4] that speeds up the ....

....all other capacities are constant, the parametric preflow algorithm finds a maximum flow and a minimum cut for each value of in an increasing sequence 1 : k . The sequence may be computed on line. The parametric preflow algorithm is a parameterized version of the Goldberg and Tarjan [5] preflowpush algorithm for computing maximum flows. The preflow push algorithm maintains at all times a 7 feasible preflow and a valid labeling. A feasible preflow is a flow f satisfying arc capacity constraints f ij u ij and relaxed node conservation constraints P i2V f ij 0 for all j 2 V ....

[Article contains additional citation context not shown here]

A. V. Goldberg and R. E. Tarjan. A new approach to the maximum flow problem. Journal of ACM, 35:921--940, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan. A New Approach to the Maximum Flow Problem. J. Assoc. Comput. Mach., 35:921--940, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan. A New Approach to the Maximum Flow Problem. J. Assoc. Comput. Mach., 35:921--940, 1988.

.... result of a similar lemma in [5] In fact, in [5] it has been shown that for some s the total number of di erent communities S i is no more than n 2 and they can all be computed in time proportional to a single max ow computation, when a variation of the Goldberg Tarjan pre ow push algorithm [7] is employed. Thus, if we want to nd a cluster of s in G of certain size or other characteristic we can simply use this methodology, nd all clusters fast and then choose the one that ts best. Also, because the parametric pre ow algorithm in [5] nds all clusters either in increasing or ....

A. V. Goldberg and R. E. Tarjan. A New Approach to the Maximum Flow Problem. J. Assoc. Comput. Mach., 35:921-940, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan, "A new approach to the maximum flow problem," J. ACM, pp. 921--940, 1988.

No context found.

Andrew V. Goldberg and Robert E. Tarjan. A new approach to the maximum-flow problem. Journal of the Association for Computing Machinery, 35(4):921--940, October 1988.

No context found.

A. Goldberg and R. Tarjan. A New Approach to the Maximum Flow Problem. Journal of the Association for Computing Machinery, 35:921--940, 1988.

No context found.

A. Goldberg and R. Tarjan, "A new approach to the maximum-flow problem," Journal of the ACM, vol. 35, pp. 921--940, 1988.

No context found.

Goldberg, A.V., R.E. Tarjan. A New Approach to the Maximum Flow Problem. J. ACM 35 (1988), 921-940.

No context found.

A. Goldberg and R. Tarjan. A New Approach to the Maximum Flow Problem. Journal of the Association for Computing Machinery, 35:921--940, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan, "A new approach to the maximum flow problem," Journal of the ACM, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan. A New Approach to the Maximum-Flow Problem. Journal of the ACM, 35(4):921--940, 1988.

No context found.

Andrew Goldberg and Robert Tarjan, "A new approach to the maximum-flow problem," Journal of the ACM (JACM), vol. 35, no. 4, pp. 921--940, 1988.

No context found.

Andrew V. Goldberg and Robert E. Tarjan. A new approach to the maximumflow problem. Journal of the ACM (JACM), pages 921--940, 1988.

No context found.

Andrew V. Goldberg and Robert E. Tarjan. A new approach to the maximum--flow problem. J. Assoc. Comput. Mach., 35(4):921--940, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan, "A new approach to the maximum flow problem," Journal of the ACM, 35(4):921-940, Oct. 1988.

No context found.

A. V. Goldberg and R. E. Tarjan. A new approach to the maximum flow problem. Journal of the ACM, 35:921--940, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan, "A new approach to the maximum flow problem," Proc. Of the 18th Annual ACM Symp. On Theory of Computing, 1987, pp. 136-146.

No context found.

A. Goldberg and R. Tarjan. A new approach to the maximum flow problem. Journal of the ACM, 1988, 921-940.

No context found.

A. V. Goldberg and R. E. Tarjan, "A new approach to the maximum flow problem," Journal of the ACM, 35(4):921-940, Oct. 1988.

No context found.

A. V. Goldberg and R. E. Tarjan. A new approach to the maximum flow problem. J. ACM, Vol. 35, No. 4, pages 921--940, October 1988.

No context found.

Goldberg,A.V. and Tarjan, R.E., A new approach to the maximum-flow problem, Journal of the ACM, 35(4):921--940, 1988.

No context found.

A. Goldberg and R. Tarjan. A new approach to the maximum flow problem. Journal of the ACM, vol. 35, 4, pages 921-- 940, 1988.

No context found.

Andrew V. Goldberg and Robert E. Tarjan. A new approach to the maximum flow problem. In Proc. of the Eighteenth Annual ACM Symposium on Theory of Computing, pages 136--146, 1986.

No context found.

A. V. Goldberg and R. E. Tarjan. A new approach to the maximum flow problem. In Proceedings of Eighteenth Annual ACM Symposium on Theory of Computing, pages 136--146, 1986.

No context found.

A. V. Goldberg and R. E. Tarjan. A new approach to the maximum flow problem. Journal of the ACM, 35:921--940, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan, "A new approach to the maximum flow problem," J. ACM, pp. 921--940, 1988.

No context found.

A.V. Goldberg and R.E. Tarjan. A new approach to the maximum-flow problem. Journal of the ACM, 35(4):921--940, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan, A new Approach to the maximum flow problem. In Proceedings of the Eighteenth Annual ACM Symposium on Theory of Computing, Journal of ACM (JACM), Vol. 35, No. 4, Oct. 1988.

No context found.

Goldberg, A. V., and Tarjan, R. E., A new approach to the maximum-flow problem. J. Assoc. Comput. Mach., Vol. 35, 1988.

No context found.

A. V. Goldberg and R. E. Tarjan. A new approach to the maximum-flow problem. Journal of ACM, 35(4):921--940, October 1988.

No context found.

Goldberg, A. V. ; Tarjan, R. E.: A New Approach to the Maximum Flow Problem. In: ACM Symposium on Theory of Computing (1986), Nr. 18, S. 136{ 146

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC