Frequency channel assignment using artificial neural networks

When planning a cellular radio network, the operator must assign frequencies or channels to each radio base station so that the probability of a given carrier to interference ratio exceeding a predefined value is sufficiently high. If the electromagnetic propagation characteristics and the spatial density of the communication traffic are known, this assignment can be reduced to an optimisation problem which is subject to interference compatibility and traffic demand constraints. The work presented investigates the problem of determining the appropriate parameters within the neural network and thus assess the feasibility of using current neural network techniques to the channel assignment problem