An efficient and robust method for solving multi-objective constraint-satisfaction problems in Cognitive Radio systems

Abstract

Cognitive radio (CR) adapts to wireless environment changes and tries to satisfy the demand of users by tuning radio parameters. However, the process of tuning the radio parameters is quite time-consuming. In order to allow a CR system to make accurate decisions, the wireless environment must be precisely modelled by reliable methods. A CR system also needs a method for tuning the radio parameters in a robust way so as to decrease the probability of doing system reconfiguration with each and every time of environment change. This work uses artificial neural network to dynamically model the environment, and proposes a method called Robust Light-weight Reasoning for Cognitive Radio that can provide robust solutions to the multi-objective problem of satisfying user given constraints.