Wireless network communication delay and channel efficiency dynamic collaborative optimization method

Abstract

The invention discloses a wireless network communication delay and channel efficiency dynamic collaborative optimization method. The method comprises the following steps: 1, establishing a wireless network communication delay and channel efficiency dynamic collaborative optimization reward model; 2, dynamically optimizing wireless network communication delay and channel efficiency reward model coefficients by using distributed reinforcement learning; and 3, completing wireless network communication delay and channel efficiency dynamic collaborative optimization control by using distributed reinforcement learning. According to the method, the wireless network communication delay and channel efficiency reward model coefficient can be dynamically optimized through reinforcement learning, so that the topological structure of the wireless network is optimized, and the wireless network communication delay and channel efficiency can be effectively and collaboratively optimized.

Description

technical field [0001] The invention relates to a method for dynamic co-optimization of wireless network communication delay and channel efficiency, belonging to the field of wireless communication. Background technique [0002] In a wireless communication network, data can be transmitted from one node to another node. In this process, the communication delay may be increased and the channel efficiency may be reduced. Both the communication delay and the channel efficiency are affected by the transmit power of the transmitting node. It is difficult for traditional methods to optimize such problems. [0003] Reinforcement learning is the process by which an agent continuously interacts with the environment to improve reward-guided actions. The goal is to optimize actions through reward settings to maximize rewards. The signal learned in reinforcement learning is an evaluation of how good or bad an action is, not how to choose the right action. Therefore, reinforcement learn...

AI Technical Summary

Problems solved by technology

However, traditional reinforcement learning cannot be single-target rewards. For this dynamic system, reward design is difficult, so it often cannot meet the requirements of optimization.

Images