Title :
On the complexity of graph clustering with bounded diameter
Author :
Jou-Ming Chang ; Jinn-Shyong Yang ; Sheng-Lung Peng
Author_Institution :
Inst. of Inf. & Decision Sci., Nat. Taipei Univ. of Bus., Taipei, Taiwan
fDate :
July 30 2014-Aug. 1 2014
Abstract :
In this paper, we study the graph clustering (partition) problem. The problem is to determine whether there is a partition of the vertices of a graph into certain number of clusters such that the diameter of subgraph induced by each cluster does not exceed a prescribed bound. An amusing result shows that for chordal graphs (respectively, dually chordal graphs) the problem is NP-complete if the diameter bound is restricted to any even integer (respectively, to any odd integer); otherwise, the problem is polynomial solvable for both classes of graphs. Moreover, by a simple reduction using graph powers, we show that there is a unified approach for solving this problem in various graph classes, including distance-hereditary graphs, doubly chordal graphs, circular-arc graphs, and AT-free graphs.
Keywords :
computational complexity; graph theory; pattern clustering; AT-free graphs; NP-complete problem; bounded diameter; circular-arc graphs; distance-hereditary graphs; dually-chordal graphs; even integer; graph classes; graph clustering complexity; graph clustering partition problem; graph powers; odd integer; polynomial solvable problem; subgraph diameter; unified approach; vertex partitioning; Approximation algorithms; Bipartite graph; Clustering algorithms; Complexity theory; Computer science; Educational institutions; Polynomials; Diamete; Graph clustering; Graph powers; NP-completeness; Special graph classes;
Conference_Titel :
Computer Science and Engineering Conference (ICSEC), 2014 International
Conference_Location :
Khon Kaen
Print_ISBN :
978-1-4799-4965-6
DOI :
10.1109/ICSEC.2014.6978122