Optimal routing scheduling method based on reinforcement learning for power wireless network

The invention discloses an optimal routing scheduling method based on reinforcement learning for a power wireless network. The optimal routing scheduling method comprises the following steps: step S101, constructing a system model by using an SDN (Software Defined Network) architecture; step S102, planning a communication link, and initializing a Q matrix; s103, the intelligent agent interacts with the environment state of the wireless communication network; step S104, updating the Q matrix; and S105, repeating the steps S103 to S104 until the network result converges, and applying the trained network to the routing selection of the wireless communication network. According to the method, channel resource allocation is reasonably adjusted by using an SDN architecture mode to meet differentiated QoS service requirements, a wireless communication network system model is established, a routing problem is mapped into a Markov decision process, iterative solution is performed, and finally, the trained network is a