User-centered distributed multi-user computing task unloading method

Abstract

The invention provides a user-centered distributed multi-user computing task unloading method. The method is characterized by comprising the following steps: uniformly abstracting multiple fog computing nodes into an access point with a servers, wherein the access point has both computing and communication capabilities; when any user generates a computing task, comprehensively considering the unloading decisions of other users; and choosing to compute locally or to unload the computing task into one of other computing nodes according to an unloading decision under a current time slot in orderto minimize own overhead. Through adoption of the user-centered distributed multi-user computing task unloading method provided by the invention, stable and efficient computing task unloading among multiple users can be realized in a distributed execution way. A user satisfactory result can be achieved, and relatively high overall performance of the system can be ensured (compared with the optimalstrategy of the system).

Description

technical field [0001] The invention relates to a user-centered distributed computing task unloading method among multiple users, which can realize stable and efficient computing task unloading among multiple users in a distributed execution manner, achieve satisfactory results for users, and ensure better overall system performance. Background technique [0002] With the rapid development of the Internet of Things, 5G communications, and artificial intelligence, a large number of new applications and services are also emerging, such as online games, ultra-high-definition video, and augmented reality, while terminal devices are showing explosive growth. On the one hand, a large amount of various data generated by these applications and devices requires a lot of resources to process; on the other hand, terminal devices generally have limited resources due to limitations in physical size and energy consumption. In order to alleviate this contradiction, mobile cloud computing ...

AI Technical Summary

Problems solved by technology

Although there have been a lot of related research work, with various metrics as optimization goals (delay, energy consumption or comprehensive consideration of delay and energy consumption), most of the research work has the following more or less deficiencies: 1) The current research work focuses on the simple scenario of multiple users, single computing node or single user, multiple computing nodes. For the more common but complex scenario of multiple users, multiple computing nodes, there is still a lack of systematic research ; 2) The current research work mostly starts from the perspective of the system, ignoring the differences in performance and goals pursued by heterogeneous users

This kind of method will cause damage to the user's own interests, thus not following the distribution results of the system, and eventually making the whole method invalid; 3) The current research work is mostly a centralized method, which is not suitable for large-scale and distributed application scenarios in the future

Images