A method to reduce network task offload delay in 6G digital twin edge computing

Abstract

The invention discloses a method for reducing the offloading delay of 6G digital twin edge computing network tasks. This paper mainly proposes a new digital twin edge computing network, which uses digital twins to estimate the state of edge servers to provide training data, and formalizes the optimization problem of a series of offloading decisions during user movement, and uses the Lyapunov optimization method to reduce the long-term migration cost The constraints are simplified to a multi-objective dynamic optimization problem, and finally the Actor-Critic-based deep learning framework is used to solve the calculation unloading optimization problem, and the agent training is realized by the digital twin edge computing network. The invention realizes the maximum reduction of unloading delay, task failure rate and migration rate under the premise of keeping low system cost.

Description

Technical field [0001] The present invention relates to the field of moving edge calculations, and specifically, a method of reducing 6G digital twin edge computing network task unloading delay. Background technique [0002] The sixth generation telecom cellular network (6G) is committed to providing better than 5G performance by digitally realizing wireless communication and calculation of real network. Moving Edge Calculation (MEC) is an important technique for implementing mobile load in 6G. With the popularization of calculation intensive and time-sensitive applications, a large number of moving Internet access devices uninstalled the task to the edge server, resulting in problems such as network delay growth and user task uninstall failure, while deploying dense markets The calculation and storage servers will increase the configuration cost. Therefore, it is especially important to optimize mobile edge computing unloading strategies. [0003] The invention of the existing d...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to use digital twins (DTs) to estimate the state of the edge server and provide a dynamic mobile offloading scheme based on deep reinforcement learning (DRL) for mobile edge computing (MEC) in view of the deficiencies in the above-mentioned prior art , to achieve the purpose of minimizing the offloading delay at the cost of accumulated service migration during user movement, and give a training framework in Digital Twin Edge Computing Network (DITEN)

Images