Partition-distance: A problem and class of perfect graphs arising in clustering

Partitioning of a set of elements into disjoint clusters is a fundamental problem that arises in many applications. Different methods produce different partitions, so it is useful to have a measure of the similarity, or distance, between two or more partitions. In this paper we examine one distance measure used in a clustering application in computational genetics. We show how to efficiently compute the distance, and how this defines a new class of perfect graphs.