Wireless distributed computing system and resource allocation method

Abstract

The invention discloses a wireless distributed computing system and a resource allocation method, which are suitable for the technical field of mobile communication. The system comprises a plurality of main nodes and computing nodes, wherein the main nodes distribute computing tasks and collect computing results, the computing nodes execute the computing tasks and feed back the computing results, and the main nodes are connected with the computing nodes through wireless channels; the resource allocation method comprises a computing node scheduling method and a computing node power allocation method; according to the channel information of each computing node, the main node selects a computing node for executing a computing task for the main node; and the computing node distributes computing power and communication power according to the channel information and the computing power thereof, so that the computing rate of the system is maximized. Data collection and joint processing work is completed by using a large number of small devices, and the requirement for a large number of special computing hardware is avoided. According to the resource allocation method, limited power can be effectively utilized, and maximization of computing power is achieved.

Description

technical field [0001] The invention relates to a wireless communication technology, in particular to a resource allocation method in a wireless distributed computing system, which belongs to the technical field of mobile communication. Background technique [0002] Group computing has attracted widespread attention in distributed sensors and distributed computing platforms. It uses a large number of wireless devices to complete data collection and joint processing, thereby avoiding the system's need for a large number of dedicated computing hardware. Taking advantage of the detachability of computing tasks, the complex computing tasks are decomposed into multiple sub-tasks that can be calculated in parallel, and distributed to different computing nodes through wireless communication. Each computing node completes a sub-computing task, and then the computing The results are transmitted to the master node for merging, forming a wireless distributed computing system. [0003]...

AI Technical Summary

Problems solved by technology

Therefore, how to reasonably allocate computing nodes and how each computing node effectively allocates its power will affect the total computing power of the entire wireless distributed computing system

Although there are studies focusing on communication aspects in wireless distributed computing systems, few works consider both communication and computing energy consumption issues

In addition, when considering the communication transmission rate, most of the work only uses the Shannon formula to calculate the single-link transmission rate, and does not consider the influence of multiple computing nodes communicating at the same time and mutual interference

Inter-user interference will seriously affect the communication rate of the system, increase communication power overhead, and reduce the computing power of the entire wireless distributed computing system

Images