Development of an information technology for network planning of an IT project based on methods of mul-tiparallel information processing
Radiotekhnika · 25.07.2026 · (225) · 104-114
DOI: 10.30837/rt.2026.2.225.10 ↗Kurz gefasst
Fachartikel in der Zeitschrift Radiotekhnika, online erschienen am 25.07.2026. Verfasst von Y.D. Tolstoluzkyi.
Abstract (Originalsprache)
The article presents a study of an information technology for network planning and operational replanning of IT projects based on the combination of probabilistic task duration estimation, temporal graph neural networks, and multiparallel data processing tools. The subject of the study is models, methods, and software-hardware tools for improving the efficiency of IT project planning while taking into account dependencies between tasks and the dynamics of project events. The aim of the work is to improve the efficiency of network planning by developing an information technology that combines a hybrid TGN + LightGBM model, a modified critical path method, and the “estimation → replanning” feedback loop within a unified software-hardware architecture. Within the study, a system of efficiency indicators was substantiated, the feasibility of using a Temporal Graph Network for predicting the remaining time of tasks was determined, the task of synthesizing a software-hardware system was formalized, and a conceptual model of the technology was developed based on the principle of alternating “synthesizer – verifier” blocks. The study is based on methods of temporal event encoding in TGN, gradient boosting prediction using hybrid features, a modified critical path method with predicted durations, and multiparallel information processing. To accelerate computations, the combination of independent operators, pipeline processing, and decomposition processing was used. The obtained results showed that the proposed technology reduces the mean absolute error over a 12-week horizon by 10.3 times compared with the classical CPM – from 288 to 28 hours. The response time to a replanning event decreased by 280 times – from 4 hours to 51.3 seconds, while the coefficient of variation of MAE decreased from more than 25% to 4.2%. A speedup coefficient of S(16) = 6.95 was achieved. The proposed technology is scalable and suitable for operational monitoring of large-scale IT projects in real time. Its advantage lies in the automatic updating of the network plan when new project events are received.
Zitieren
APA
Tolstoluzkyi, Y. (2026). Development of an information technology for network planning of an IT project based on methods of mul-tiparallel information processing. Radiotekhnika, 104-114. https://doi.org/10.30837/rt.2026.2.225.10
Vancouver
Tolstoluzkyi Y. Development of an information technology for network planning of an IT project based on methods of mul-tiparallel information processing. Radiotekhnika. 2026:104-114. doi:10.30837/rt.2026.2.225.10
BibTeX
@article{Tolstoluzkyi2026,
author = {Tolstoluzkyi, Y.D.},
title = {Development of an information technology for network planning of an IT project based on methods of mul-tiparallel information processing},
journal = {Radiotekhnika},
year = {2026},
number = {225},
pages = {104--114},
doi = {10.30837/rt.2026.2.225.10}
}
EUVIMED