Risk sensitive resource control approach for delay limited traffic in wireless networks

In this paper, we consider the challenging problem of power allocation in interfering wireless networks with services having stringent delay constraints. The objective is to allocate optimally the power to each user, take advantage of the time diversity of the time varying channel and satisfy the delay constraints of the services. To this end, we model the bit rate dynamics of the users as a linear stochastic control scheme and formulate a risk-sensitive control framework that minimizes an exponential cost function of the deviation between the traffic arrivals and departures from the queues. We provide the optimal solution of the power control problem and provide simulation results to assess the superiority of our solution compared to some existing solutions.