Partitioning WCN Models for Parallel Simulation of Radio Resource Management

Parallel Simulation techniques have been proposed as a possible solution to execution time and memory constraints often found in detailed simulations of Wireless Cellular Networks. However, partitioning represents a major challenge for models that encompass elements of radio propagation phenomena. This paper discusses the partitioning problem with respect to Parallel Discrete Event Simulation and we formulate an approach to study partitioning of a WCN model that includes radio propagation. Various options for a model of moderate size and where interference is calculated over the whole system are evaluated through experimentation and some limited mathematical analysis. Results indicate that radio spectrum based partitionings is preferable to geographically based partitionings for this model in many realistic scenarios. It is also noted that characteristics of the model differ sufficiently from other previously studied spatially explicit problems to reduce or even annihilate the effectiveness of some commonly used partitioning techniques.