A distributed radio resource usage optimization algorithm to satisfy QoS requirements by neural network dynamics

We propose a distributed radio resource usage optimization algorithm to satisfy users' QoS requirements in heterogeneous wireless networks. Our optimization algorithm is based on the mutually connected neural network dynamics, which can be run distributively and autonomously, and does not require centralized computation. We define an optimization problem to satisfy QoS requirements of users' applications, and compose a neural network to solve the problem autonomously. Since the objective function of the defined problem becomes higher order function, we apply a higher order neural network model that minimizes higher order energy function by its distributed and autonomous updating. By computer simulation, the proposed method has been shown effective to optimize such a problem, without any centralized computation.