DocumentCode :
288997
Title :
A parallel local-search algorithm for the k-partitioning problem
Author :
Diekmann, Ralf ; Lüling, Reinhard ; Monien, Burkhard ; Spraner, C.
Author_Institution :
Dept. of Math. & Comput. Sci., Paderborn Univ., Germany
Volume :
2
fYear :
1995
fDate :
3-6 Jan 1995
Firstpage :
41
Abstract :
Presents a new algorithm for the k-partitioning problem which achieves an improved solution quality compared to known heuristics. We apply the principle of so-called “helpful sets”, which has been shown to be very efficient for graph bisection, to the direct k-partitioning problem. The principle is extended in several ways. We introduce a new abstraction technique which shrinks the graph during runtime in a dynamic way, leading to shorter computation times and improved solution qualities. The use of stochastic methods provides further improvements in terms of solution quality. Additionally, we present a parallel implementation of the new heuristic. The parallel algorithm delivers the same solution quality as the sequential one, while providing reasonable parallel efficiency on moderately sized MIMD systems. All results are verified by experiments for various graphs and processor numbers
Keywords :
computational complexity; graph theory; heuristic programming; parallel algorithms; search problems; MIMD systems; abstraction technique; computation times; graph bisection; helpful sets; heuristics; k-partitioning problem; parallel efficiency; parallel local-search algorithm; processor numbers; runtime dynamic graph shrinking; solution quality; stochastic methods; Computer science; Humans; Mathematics; Multiprocessor interconnection networks; Parallel architectures; Parallel processing; Partitioning algorithms; Runtime; Solid modeling; Stochastic processes;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
System Sciences, 1995. Proceedings of the Twenty-Eighth Hawaii International Conference on
Conference_Location :
Wailea, HI
Print_ISBN :
0-8186-6930-6
Type :
conf
DOI :
10.1109/HICSS.1995.375478
Filename :
375478
Link To Document :
بازگشت