Hypergraph models for cellular mobile communication systems

Cellular systems have hitherto been modeled mostly by graphs for the purpose of channel assignment. However, hypergraph modeling of cellular systems offers a significant advantage over graph modeling in terms of the total traffic carried by the system. For example, we show that a 37-cell system when modeled by a hypergraph carries around 30% more traffic than when modeled by a graph. We study the performance of channelized cellular systems modeled by hypergraphs in comparison with those modeled by graphs. For this purpose, we have evaluated the capacities of these cellular networks defined in McEliece and Sivarajan (1994). Evaluation of the capacity necessitates generation of maximal independent sets of hypergraphs. We describe some new algorithms that we have developed for this purpose.