Abstract
This article carries out multi-objective optimization for the packet loss rate and delay of the Bluetooth Mesh network. Apply GRU-based network flow prediction and DDPG's intelligent routing method, and obtain the Bluetooth Mesh network load information under the SDN architecture to construct a Mesh network including delay, using the original traffic matrix and the predicted traffic matrix as the state space, based on deep determinism Policy Gradient (DDPG) algorithm simulation to obtain optimal routing, effectively allocate routing resources and manage network traffic. Experimental results show that compared with before optimization, the routes generated by this algorithm can reduce the packet loss rate and delay of the Bluetooth Mesh network.