Topology Optimization of Wireless Sensor Networks based on Full Interactive Distributed Reinforcement Learning

Deploying a large number of sensor nodes within a certain area can form Wireless Sensor Networks. Wireless Sensor Networks enable real-time collection and monitoring of environmental information through communication between sensor nodes and adjust the number of sensor nodes to meet the requirements of different application scenarios. However, Wireless Sensor Networks are susceptible to environmental influences. Dynamic changes in the environment may lead to a decrease in communication quality or even interruptions of communication links, resulting in changes in the network's topology structure. This paper proposes a topology optimization method for WSNs based on Fully Interactive Distributed Reinforcement Learning. The algorithm equips the agent with the ability to assess correlations between sensor nodes and update its parameters by obtaining global parameter information, which benefits its learning process. Subsequently, the agent devises a strategy based on the updated parameters. Through this distributed reinforcement learning method, the transmission power of sensor nodes can be dynamically and efficiently adjusted, providing a low-latency and highly stable communication network for wireless sensor networks.