Near-Optimal Working Slots Selection for a Sensor Via Submodularity
Xiaopei Wu, Yue Wu · 2009
In this paper, we consider the sensor working periodically and every period was divided into N slots. To reduce energy consumption, the sensor can only be active in K slots of N for every period, the basic problem is which K slots are fine enough (like: minimum packet delay). According to the requirement of packet delay, we introduce a monotonic non-decreasing expected penalty reduction function as our criterion to select working slots, which is a submodular set function. By exploiting the submodularity property, we give a simple greedy algorithm, which is a constant (1 - 1/e) of the optimal working slots set. The experimental results show that the proposed algorithm can find the optimal working slots for nearly all energy conserving rate.