Exploiting Different Cognition Levels of Channel Information in Transmission Scheduling of Two Sources Over Time Varying Wireless Channels

In this work, we investigate the methods to optimally exploit different cognition levels of channel information in transmission control for interfering sources. We consider a network that consists of two transmitter-receiver pairs which operate over time varying channels. We obtain the optimal scheduling techniques which maximize the expected weighted sum-rate of the network per time slot. The scheduling decisions depend on the information about the channels between nodes. Due to energy and timing overhead, learning instantaneous channels states may be costly or infeasible. We consider the optimal scheduling policies in the following cases: (1) when all the channels are perfectly measured, (2) when all the channels are measured with some error probability, (3) when delayed information about the channels is known using previous time slots transmissions, (4) when the channels are measured infrequently, (5) when the channels are not measured at all, or (6) when the decisions are taken in a distributed manner. In each of the cases, we exploit the channel characteristics to formulate the scheduling optimization problem. We derive the formulas for the belief about the channels to be in a certain state at a certain time slot. We formulate the problem of finding the optimal policy in the case of availability of delayed information as a partially observable Markovian decision problem (POMDP). In the case of distributed scheduling, we formulate the problem as a quadratic program. We compare the optimal performance levels in all the cases numerically and thus evaluate the effect of channel state information knowledge.