An energy function approach and its application to a new clustering algorithm

The paper addresses the problem of natural partitioning of a set of geographically distributed nodes. For this purpose an energy function has been defined. Based on this energy function a clustering algorithm, which partitions the nodes into clusters, has been presented. It has been assumed, intuitively, that if the energy function attains a peak, the clusters produced follow the natural partitioning. The algorithm has been tested on a set of non-trivial node distribution.<>