The Impact of Imperfect Information in Multi-Channel Wireless Systems

We consider a wireless system with multiple channels when each channel has several different transmission states associated with different probabilities of successful transmissions. Hardware limitations and MAC protocols dictate that the channel be switched only after certain intervals, and the channel needs to be selected based on the channel states at the beginning of the interval. We demonstrate that the fundamental relations between QoS metrics like throughput and stability significantly change owing to the lack of complete information. We obtain a randomized channel selection and threshold-type transmission rule that maximizes throughput while using imperfect information. Using this optimal strategy, we numerically quantify the penalty associated with different amounts of imperfections in the available information.