AUTHOR=Sun Maojin , Sun Luyi 

TITLE=Optimization of artificial intelligence in localized big data real-time query processing task scheduling algorithm

JOURNAL=Frontiers in Physics

VOLUME=Volume 12 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2024.1484115

DOI=10.3389/fphy.2024.1484115

ISSN=2296-424X

ABSTRACT=The development of science and technology has driven rapid changes in the social environment. The progress of various science and technologies has made people's lives more and more convenient. The most obvious is today's big data environment. In the big data environment, people are getting faster and faster in the search for information, which is inseparable from the design of computer systems. In the process of processing big data, it is necessary to divide and allocate the information and resources of big data reasonably. However, there are often problems such as unreasonable allocation of information resources, so task scheduling algorithms are used to process big data. The task scheduling algorithm itself is more dependent on the computer system and the algorithm program. In the actual processing process, the performance differences of different nodes are hardly considered. Therefore, the optimization of the task scheduling algorithm is a problem that needs to be considered at present. In this paper, the task scheduling algorithm is optimized by the method of artificial intelligence, and the optimization model of the task scheduling algorithm is designed by using support vector machine, K-NearestNeighbor, and fuzzy comprehensive evaluation. The experimental results are as follows:
1. The algorithm used by the research has better classification accuracy on the Mashroom, Chess, Sonar, and Wine datasets, with an average accuracy of 87.23%. Next is the backpropagation neural network, with an average classification accuracy of 84.83%.
2. The optimized model significantly reduces task processing time and total cost, with a maximum reduction of 2935ms in task processing time.
3. The optimized scheduling scheme significantly reduces the query time for the same statement, with the shortest time being only 0.83s, which is 1.13s less than the pre-optimized scheme.
The experimental results all prove that the optimization model in this paper has certain practicability.