Optimal Sleep Scheduling for a Wireless Sensor Network Node

We consider the problem of conserving energy in a single node in a wireless sensor network by turning off the node´s radio for periods of a fixed time length. While packets may continue to arrive at the node´s buffer during the sleep periods, the node cannot transmit them until it wakes up. The objective is to design sleep control laws that minimize the expected value of a cost function representing both energy consumption costs and holding costs for backlogged packets. We consider a discrete time system with a Bernoulli arrival process. In this setting, we characterize optimal control laws under the finite horizon expected cost and infinite horizon expected average cost criteria.