文摘
Reduction of unnecessary energy consumption is becoming a major concern in green wireless sensor networks. Sleep scheduling is one of the efficient strategies to achieve energy saving. In this paper, we propose a novel scheme for the sleep scheduling, which is based on Decentralized Partially Observable Markov Decision Process (Dec-POMDP). A sleep scheduling algorithm with online planning (Dec-POP-SSA) with respect to Dec-POMDP is also presented. In Dec-POMDP, due to the hardness of obtaining the state spaces and the reward with mold-free environment, quasi-Monte Carlo is applied to collect state spaces such that the real-time acquisition of beliefs state is achieved, and the reward is evaluated in tracking reward and coverage connectivity intensity. Instead of producing the entire plan, Dec-POP-SSA need only find actions for the current step. We also give the theoretical analysis on the upper bound for Dec-POP-SSA. The numerical experiments show that Dec-POP-SSA may receive the highest reward.