Optimization for Time Consumption on Channel Searching in Cognitive Radio System with Markov Decision Processes

As the demand for spectrum resources becomes more and more strong, and some spectrum has been underutilized for years, the concept "Cognitive Radio" becomes popular and is widely applied in practice. In a cognitive radio system, one of the important capabilities of cognitive users is that, they should or we may say they "have to" be able to search for a better channel to keep the users' data on transfer with acceptable quality when the interference in the current channel exceeds the limit they can live with. But practically the procedure of channel searching takes time and too much time consumption on that will definitely delay users' data transfer so that the system performance will get degraded, this is what a cognitive user has to be confronted with inevitably. From this perspective, as far as a cognitive user is concerned, how to decrease time consumption for channel searching as possible as we can becomes a critical technical problem. In this paper, we propose that this problem be modeled as a Markov decision processes (which is also referred to as MDP throughout this paper) such that we can solve this problem by the existing algorithms derived from the previous research on Markov decision processes and the results show that by using the algorithm of policy improvement, the long-term average time consumption for channel searching can be effectively minimized.