The static channel assignment (SCA) problem arises in cellular telephone networks where discrete frequency ranges within the available radio frequency spectrum, called channels, need to be allocated to different geographical regions in order to minimise the overall interference.

Assigning channels to regions in order to minimise the interference generated has been shown to be a graph coloring problem and is therefore NP-hard. Many techniques for finding solutions to the SCA problem have been tried including simulated annealing, neural networks, and genetic algorithms.

The parameters for the genetic algorithm approach need to carefully chosen in order to cause the algorithm to rapidly converge to a suitably low minimum. Finding appropriate values for these parameters can be a time consuming process because the algorithm must be executed many times with a range of data sets, mutation rates and population sizes.

By distributing the workload over a number of workstations using Clustor, a wide range of parameter values can be tested without the need for continuous human intervention, and in a fraction of the time it would take on a single computer.

In this study two parameters were explored:

• The mutation rate
• The population size