Attention mechanism reinforcement learning-based edge network device caching method

Abstract

The invention discloses an attention mechanism reinforcement learning-based edge network device caching method, which comprises the following steps that a cellular network model is established, the cellular network model comprises user equipment, edge network devices and a core network, and each edge network device is internally provided with an action network module and an evaluation network module; the edge network devices receive a request sent by the user equipment in the area where the edge network device is located; each edge network device obtains observed values of other edge network devices; each edge network device selects an action according to a cache replacement strategy and an observation value; the edge network device sends the action and the updated state to the adjacent edge network device; parameters of the action network module and the evaluation network module are updated according to the updated observation value of the action and the action value function; and the cache replacement strategy is optimized according to the target function. According to the invention, the frequency of repeatedly downloading the file from the cloud data center can be reduced, the delay is reduced, and the network service quality and the user experience quality are improved.

Description

technical field [0001] The invention relates to the technical field of edge caching and deep reinforcement learning, in particular to an edge network device caching method based on attention mechanism reinforcement learning. Background technique [0002] With the development of network technology and the surge of demand, the speed and throughput of data and applications are causing the rapid growth of traffic. This challenge has also promoted the urgent revolution of network architecture and advanced communication technology. Mobile Edge Computing (MEC) technology can effectively alleviate the traffic pressure of mobile network operators. By storing content on base stations or local devices close to users, it can effectively reduce redundant data transmission delays in application services and improve service quality. [0003] In real life, users are in various scenarios. Due to the different services provided in different scenarios, the content cached by each base station...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the technical problem that the traditional cache replacement strategy does not consider dynamic interaction, this invention proposes an edge network device caching method based on attention mechanism reinforcement learning, which solves the multi-agent edge problem by introducing the attention mechanism into the actor-evaluator algorithm. caching problem

Images