Graphics processing unit (GPU) parallel particle swarm optimization (PSO) method based on Amason web service (AWS)

Abstract

The invention discloses a graphics processing unit (GPU) parallel particle swarm optimization (PSO) method based on an Amason web service (AWS). The method comprises the following steps: initializing the position and velocity of each particle of an AWS host side; initializing an individual best position and a global best position of the AWS host side; updating the individual position and velocity of an AWS equipment side in parallel; computing objective function value of the AWS equipment side; and updating the individual best value and the global best value of an AWS equipment side in parallel, and continuously calculating after evaluating whether the values are met the standard. According to the technical scheme of the GPU parallel PSO method based on AWS, an AWS cloud computing platform is used to accelerate paralleling of the PSO under large-scale data, the technology brings the cloud computing platform into small production environments and university research laboratories, and guides the small production environments and the university research laboratories to use massive computing resources in a cloud computing age.

Description

technical field [0001] The invention relates to an AWS-based GPU parallel particle swarm algorithm. Involves patent classification number G06 Calculation; Calculation; Counting G06N Computer system based on specific calculation model G06N3 / 00 Computer system based on biological model G06N3 / 02 Using neural network model. Background technique [0002] Particle swarm optimization (PSO) originated from the study of bird predation behavior, and is an evolutionary computing technology based on swarm intelligence. It shows great potential in practical engineering. However, in many fields such as numerical modeling and optimization calculations, when dealing with large amounts of data and solving large-scale complex problems, the PSO algorithm still requires a lot of computing time. [0003] The iterative speed of a computer graphics processing unit (Graphics processing unit, GPU) exceeds that of a CPU in the same period. Since GPU has a very high memory bandwidth and a large num...

AI Technical Summary

Problems solved by technology

[0004] However, the existing GPU-based PSO parallel acceleration only utilizes the function of the local single-node GPU, and cannot utilize the big data processing capability of the GPU device; GPU hardware devices, especially GPU cluster devices with big data processing capabilities, are expensive and expensive Research institutions cannot afford it; apply the cloud computing platform AWS (Amason Web Service) to the field of GPU parallel computing to control the cost of research and production

Images