Optimization Method of Opportunistic Network Routing Based on Deep Reinforcement Learning
Department of Information and Intelligent Engineering, Ningbo City College of Vocational Technology, Ningbo 315199 School of Graduate Studies, Management and Science University, Shah Alam, Malaysia
Md Gapar Md Johar
Software Engineering and Digital Innovation Centre, Management and Science University, Shah Alam, Malaysia
Ali Khatibi
School of Graduate Studies, Management and Science University, Shah Alam, Malaysia
Jacquline Tham
School of Graduate Studies, Management and Science University, Shah Alam, Malaysia
Abstract
The conventional routing optimization method of opportunistic networks mainly solves the problem of network breakage, which can not meet the communication conditions of nodes in complex environment. Therefore, an opportunistic network routing optimization method based on deep reinforcement learning is designed. Optimize the routing mechanism of opportunistic network, use routing protocol to obtain the next node, and transmit network data with three hop paths, and control the communication overhead ratio by reducing the number of data copies, thus ensuring the data transmission efficiency. Based on deep reinforcement learning, the opportunistic network congestion-aware probabilistic routing protocol is selected, and the redundancy of packet forwarding and caching is managed according to the contact probability of opportunistic network nodes and the action value of packet state, thus alleviating the congestion problem of long-term storage nodes of packets. The simulation results show that the optimization effect of this method is better and can be applied to real life.