Power Allocation in a Wireless Network Using Scheduling and Stochastic Programming with Penalty Functions

This paper studies the optimal allocation of transmit power in a wireless communication network. First, a stochastic programming formulation is introduced, based on penalizing violations of quality-of-service constraints. The maximization of the certainty equivalent signal-to-interference ratio under Rayleigh fading corresponds to a penalty model where (max-min) fairness is explicitly taken into consideration. Second, optimum dynamic power allocation is discussed. Efficient dynamic resource allocation under both linear and logarithmic utility functions is addressed. The dynamic model studies the optimal trade-off between instantaneous quality-of-service and a delay-penalized reliable quality-of-service. Related work on optimal stochastic power control is summarized.