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.