Abstract: "Parallel algorithms for special cases of the assignment problem have been designed. These algorithms assume the edge weights are integers and within a range. In one case the algorithm is good if the maximum of the absolute values of the edge weights is polynomial in the number of ver...

Ausführliche Beschreibung

Bibliographische Detailangaben

Hauptverfasser:	Osiakwan, Constantine N. (VerfasserIn), Akl, Selim G. (VerfasserIn)
Format:	Buch
Sprache:	English
Veröffentlicht:	Kingston, Ontario, Canada 1989
Schriftenreihe:	Queen's University <Kingston, Ontario> / Department of Computing and Information Science: Technical report 258
Schlagworte:	Algorithms Graph theory Parallel processing (Electronic computers)
Zusammenfassung:	Abstract: "Parallel algorithms for special cases of the assignment problem have been designed. These algorithms assume the edge weights are integers and within a range. In one case the algorithm is good if the maximum of the absolute values of the edge weights is polynomial in the number of verticies, n. In another case the time-processor product exceeds the running time for best sequential algorithm for the assignment problem In this paper, we present an adaptive parallel algorithm for the assignment problem of a complete weighted bipartite graph, where the edge weights can be real valued. The algorithm is designed using the exclusive-read exclusive-write parallel random-access machine (EREW PRAM) model of parallel computation. For a complete weighted bipartite graph of n verticies, the algorithm runs in O(n[superscript 3]/p + n[superscript 2]p) time using p processors. We obtain a perfect speedup, with respect to the O(n[superscript 3]) Hungarian method, for p [lesser than or equal to the square root of n].
Beschreibung:	43 S.