Proximity graph based clustering algorithms for optimized planning of UMTS access network topologies

Abstract

After a short overview of UMTS radio access networks and their planning process, two new clustering approaches for partitioning the set of given Node B into a near optimum number of RNC-clusters are presented. Both algorithms start from a proximity graph, consider all relevant cost contributions, and provide flexibility by use of user-supplied cost functions for equipment and links. Their high speed allows the embedding into an interactive tool environment and leaves room for more time consuming local search procedures for further improvement of the solutions.