Social perception D2D collaborative caching method based on deep Q learning

Abstract

The invention discloses a social perception D2D collaborative caching method based on deep Q learning, and aims to solve the problem of link congestion caused by redundant content transmission in a network. The method comprises the following steps: step 1, initializing a communication scene; 2, performing multi-attribute modeling on the social relation strength of the equipment owner; step 3, modeling a D2D cooperative caching process based on a reinforcement learning process; step 4, designing a network structure of a Behavior network for training and a Target network for calculation in the deep Q network; and step 5, based on a set system model, utilizing a deep reinforcement learning algorithm to enable the intelligent agent and the environment to interact to carry out iterative learning to output an optimal strategy, namely, an optimal decision of equipment cache resource allocation and D2D content distribution. According to the technology, normal communication of cellular users can be ensured, the average time delay of the system can be minimized on the premise of meeting personal willingness of the users, and the satisfaction degree of the users is improved to the maximum extent.

Description

technical field [0001] The present invention relates to the field of communication technology, in particular to a socially aware D2D collaborative caching method based on deep Q-learning. Background technique [0002] With the rapid development of mobile communication technology and the massive access of user equipment, a large number of calculation-intensive and delay-sensitive mobile application computing and communication tasks have caused the explosive growth of mobile data traffic in the network, which leads to the The backhaul links between them are heavily burdened. It is estimated that by 2022, nearly half of the devices and connections on the Internet will have video capabilities, and multimedia data will account for more than 82% of total IP traffic by then. Studies have confirmed that users' requests for content on the Internet are often concentrated, and requests for 20% of popular content account for 80% of network traffic, resulting in repeated and intensive c...

AI Technical Summary

Problems solved by technology

Due to the complexity of communication scenarios and the uncertainty of user service requests, the current reinforcement learning method based on Q-learning will face severe challenges such as large state-action space, poor reinforcement learning performance, slow convergence speed, and unstable learning results. question

As a result, the system has a high blocking rate and low throughput, making it difficult to meet actual communication needs.

Images