Dynamic Cell Clustering in Cellular Multi-Hop Networks

This paper proposes, examines and compares different algorithms for dynamic cell clustering in cellular multi-hop networks. The basic idea is to balance load among different cells by creating cell clusters dynamically. For that purpose, it is distinguished between centralized and decentralized heuristics, which both are compared with results of a numerical optimization. At first, the evaluation is performed based upon a graph-theoretical network model. In a second step, an implementation of the decentralized approach in ns-2 is used to analyze and prove the protocol´s behavior and feasibility in a more realistic environment