基于强化学习的水下传感网负载均衡路由算法  

Load balancing routing algorithm of underwater sensor network based on reinforcement learning

作　　者：叶晓涵 李德识[1] YE Xiao-han;LI De-shi(Electronic Information School,Wuhan University,Wuhan 430072,China)

机构地区：[1]武汉大学电子信息学院,湖北武汉430072

出　　处：《计算机工程与设计》2020年第12期3301-3307,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61571334)。

摘　　要：为解决负载分配不均衡会限制水下无线传感网中吞吐量等性能的问题,提出一种基于强化学习的分布式水下负载均衡路由算法,算法中水下节点根据对应父节点的带宽状态,分布式地学习分配收益并进行负载分配。通过强化学习构建带宽状态和负载分配决策之间的关系模型,使节点在不同带宽状态下进行分配决策,引入演化博弈论来优化决策选择的策略,加快学习过程的收敛速度。仿真结果表明,该算法可实现分布式路由选择和网络负载的均衡分配,有效改善了网络性能。To solve the problem that imbalanced load allocation limits throughput and other performance in underwater wireless sensor networks,a distributed underwater load balancing routing algorithm based on reinforcement learning was proposed.In the algorithm,the nodes learnt distributes benefits in an allocation manner and allocated load according to the bandwidth of their parent nodes.Reinforcement learning was used to build a relationship model between bandwidth states and load allocation decisions,nodes made allocation decisions under different bandwidth states.Evolutionary game theory was introduced to optimize decision-making strategies to speed up the convergence of the learning process.Simulation results show that the proposed algorithm can realize distributed routing and network load allocation,and effectively improve network performance.

关 键 词：水下无线传感网 负载均衡 分布式决策 强化学习 演化博弈论 

分 类 号：TN929.3[电子电信—通信与信息系统]