With the development of the Internet of things industry in the 21st century,wireless sensor networks(WSN)as one of its core technologies,the relay and forwarding efficiency of its networking nodes to sensor information is particularly important in various application scenarios.The advantages and disadvantages of network routing,traffic allocation and node energy planning will directly affect the overall operation efficiency of the network and the resulting economic benefits.In recent years,the popular deep reinforcement learning algorithm is outstanding in decision-making optimization.In this study,it is applied to the routing strategy of wireless sensor network nodes in the form of distributed microservices,giving full play to the computing and storage resources of each node in the network to cooperate to complete the overall routing decision of the network,It effectively improves the shortcomings of traditional centralized routing strategy in the face of large-scale wireless sensor networks,such as large sensing information delay and inconvenient expansion.This paper studies the specific application scenario of wireless sensor networks,and proposes a routing optimization scheme based on deep reinforcement learning algorithm.By perceiving and comprehensively considering the environment state,the service measurement model of nodes and links is analyzed to provide more effective environment parameters for the training of deep reinforcement learning algorithm,so as to improve the efficiency and accuracy of the optimal routing algorithm.The proposed two hop node awareness strategy balances the perceived range of nodes and the timeliness of perceived information,and fully excavates the network service capability.This scheme not only makes the wireless sensor network more intelligent in the dynamic deployment environment,but also improves the overall lifetime and survivability of the network.In this paper,pytorch framework is used to complete the construction and training of deep reinforcement learning model,and ddqn model is generated.The trained model is simulated with NS-3.Compared with the existing algorithms to analyze the advantages and disadvantages of the proposed routing algorithm,the results show that the proposed method is consistent with the theoretical analysis,and has certain feasibility and good performance. |