A parallel algorithm for maximum independent set of a circular-arc graph

Abstract

The author presents a parallel algorithm of cost O(n log n) to find a maximum independent set of a circular arc graph. In the CREW PRAM model the algorithm takes O(log n) time, while in the EREW PRAM model it requires O(log/sup 2/ n) time. It illustrates the use of divide-and-conquer in parallel algorithms. The heart of the algorithm solves this problem on an interval graph, which is derived from the given circular arc graph. Postprocessing selects a maximum independent set on the given circular arc graph.<>