A Cloud-Edge Collaborative Computing Migration Method Based on Deep Reinforcement Learning

Abstract

The invention discloses a cloud-edge collaborative computing migration method based on deep reinforcement learning, comprising the following steps: (1) constructing a mobile edge network computing migration model composed of a user layer, an edge cloud layer, and a cloud layer; (2) building a mobile edge network computing migration model in the edge cloud layer The edge cloud node receives the computing task of user layer migration, and allocates broadband and computing resources according to the task; (3) takes the minimum delay and energy consumption of the computing task as the optimization goal, and constructs the objective function and constraints; (4) constructs the deep neural network The model uses the asynchronous edge-cloud collaborative deep reinforcement learning method to optimize the objective function, obtains the optimal migration decision, and returns to the terminal at the user layer to execute the decision. The present invention can solve the problem of solving complex combination optimization in big data scenarios, solves the disadvantage of slow convergence speed in traditional gradient optimization methods, and the method consumes less time when processing large-scale data, and can adapt to the timeliness requirements of big data network data processing .

Description

technical field [0001] The invention relates to a calculation migration method, in particular to a cloud-edge collaborative computing migration method based on deep reinforcement learning. Background technique [0002] In the cloud computing model, a large amount of data perceived by terminal devices will be migrated to a centralized cloud server for processing, thereby greatly expanding the computing power of terminal devices. However, in the cloud computing model, the cloud server is usually far away from the terminal device, resulting in high latency of application task processing and high overhead of remote transmission of large-scale data. The existence of such problems has given birth to the edge computing model. As a new computing model, edge computing sinks computing and storage resources to the edge of the access network close to terminal devices, trying to integrate cloud service providers, mobile operators Provides deep integration with heterogeneous Internet ter...

AI Technical Summary

Problems solved by technology

[0004] The first type of schemes are basically based on exact methods or approximate methods based on mathematical programming to solve the corresponding computational migration optimization problems. It is difficult to solve complex combinatorial optimization problems in big data scenarios, and the solution methods are difficult to make according to the corresponding actual scene changes. Adaptive Migration Decisions

[0005] The second type of scheme combines computational migration research of machine learning theory, and most of them use a single deep learning or reinforcement learning theory to solve corresponding optimization problems. This type of solution method fails to give full play to the advantages of machine learning in perception and decision-making , so that the solution is limited

The above-mentioned solution methods based on deep reinforcement learning, when faced with a scenario with too many edge nodes or a huge amount of tasks, due to the limitations of resources in the edge cloud and the correlation of processing tasks between edge nodes, such methods often cannot be based on real-time Differences in dynamic environments make optimal migration decisions. At the same time, more and more migration decisions are stored in shared memory, which will lead to an excessive storage burden on edge nodes and affect the processing efficiency of edge nodes.

Images