Scheduling is a mature, but unresolved, problem in wireless systems that share the same channel. It amounts to determining when each user transmits and at what rate. The answer depends on many parameters of the system, like channel characteristics, detector structure, transmission power, modulation, coding, etc but also on user demand. It is known that multiple users can transmit simultaneously and successfully if they adjust their transmission rates properly.
The version of the problem that we consider is based on the fact that to minimize latency and overall delay it suffices to minimize “service time”. Thus we focus on a version of the problem where the demand is expressed in terms of fixed “bit”-volumes that reside at the transmitter buffers and need to be transmitted in minimum time. This is also known as the “empty-the-network” problem. It is possible to formulate the problem as a linear program where the variables are the activation durations of the user sub-sets and the coefficients are the corresponding transmission rates. In addition to solution characterization, complexity and structural results, and efficient solution methods we are imbedding this work in a broader optimization context that combines multi-user information theory with optimization methodologies. Furthermore, we have developed practical algorithms and approximations that permit the application of our theoretical investigation to practical systems. Read More