A dynamic channel allocation technique based on Hopfield neuralnetworks

摘要

The interest in global spectrum allocation techniques is growingnwith the always increasing spectrum demand for (cellular) mobilencommunications. However, the best algorithms suffer from highncomputational times that reduce the possibility of a practicalnimplementation. This paper deals with a dynamic channel allocation (DCA)ntechnique based on an energy function whose minimization gives thenoptimal allocation. Due to the particular formulation of such an energynfunction, the minimization can be performed by a Hopfield neural networknfor which a fast hardware implementation has been recently proposed innthe literature. The performance of the proposed DCA technique is derivednby computer simulations. Comparisons with a classical fixed allocationntechnique (FCA) and a different DCA technique are shown to highlight thenbetter performance of the proposed DCA technique