Development of decentralized artificial intelligence algorithms for data processing on the Internet of Things based on graph databases

Abstract

This article presents a comprehensive study on the development of decentralized artificial intelligence algorithms for data processing on the Internet of Things (IoT) using graph databases. The relevance of the work is determined by the growing volume of information generated by IoT devices and the need to create fault-tolerant systems capable of efficiently processing complex data relationships without centralized control. The Introduction section substantiates the problem of traditional centralized systems that are vulnerable to failures and maladaptive to scaling, which requires the use of decentralized approaches that integrate AI algorithms and graph structures. The developed methodology is described further in the "Methods" section.: the choice of a decentralized algorithm architecture, the use of graph databases to model connections between IoT devices, and the introduction of machine learning algorithms for predictive analysis and information flow processing. The experimental platform was simulation models of the IoT network, which provided verification of the operability of the proposed algorithms in real conditions. The "Results" section demonstrates that the proposed architecture provides a significant improvement in system parameters.: reducing response time, increasing fault tolerance, and optimizing the allocation of computing resources through a decentralized approach. The experiments have confirmed that the adaptation of AI algorithms to work with graph structures makes it possible to effectively identify hidden dependencies between devices, optimizing the processes of integration and data analysis. The "Discussion" section analyzes the advantages and limitations of the algorithms under study. The issues of scalability, security, and potential integration with existing distributed systems are discussed, which opens up prospects for further research in this area. The authors emphasize that the proposed approach contributes to the development of new models of device interaction in the Internet of Things, capable of adapting to dynamic changes in the network. Thus, the presented research makes a significant contribution to the development of IoT data processing technologies by providing reliable and scalable solutions capable of meeting the needs of modern distributed information systems.