基于强化学习的IEEE 802.15.4网络区分服务策略

为了弥补IEEE 802.15.4协议原有区分服务机制的不足，提出了一种基于BCS(backoff counter scheme)与强化学习的区分服务策略。从终端节点出发，在原优先级区分服务策略的基础上增加BCS退避策略以解决流量较大场合业务区分问题；针对协调器节点，提出了基于强化学习的占空比调整策略，该策略能根据不同应用需求和环境变化自适应调整占空比。仿真结果表明，提出算法能针对不同环境满足高优先级业务性能需求，并能根据流量变化进行占空比调整，具有极强环境适应性。

IEEE 802.15.4 differentiated service strategy based on reinforcement-learning

To provide better support in differentiated service for IEEE 802.15.4,a novel differentiated service mechanism was proposed based on BCS(back off counter scheme)and reinforcement learning.In terms of end-device,BCS backoff strategy was added to original priority-based differentiated strategy to solve the service differentiation problem under higher traffic condition.Whil “ e for the coordinator,a reinforcement learning based duty-cycle adjustment algorithm was proposed toself-learning”an optimal duty-cycle according to different application requirements and environmental changes.Simulation shows that the proposed algorithm can meet the performance requirements of high-priority service under different environments and adjust the duty-cycle when traffic is changed,which showed a strong environmental adaptability.