LIMITATIONS OF RESOURCES OF MOBILE AND EDGE PLATFORMS WHEN DEVELOPING SOFTWARE FOR IMPLEMENTATION AND OPTIMIZATION OF MACHINE LEARNING MODELS

Authors

DOI:

https://doi.org/10.28925/2663-4023.2026.34.1360

Keywords:

resource constraints, artificial intelligence, machine learning, software development,, mobile platforms, edge platforms

Abstract

This paper analyzes modern approaches to the implementation and optimization of machine learning models for mobile and edge platforms. The relevance of this issue is due to the rapid development of mobile computing, Internet of Things systems, peripheral computing, and the need to perform intelligent data processing directly on devices with limited resources. It is shown that modern mobile and edge devices are gradually transforming into full-fledged computing systems that are capable of performing complex machine learning and artificial intelligence tasks without the use of remote servers. At the same time, the use of models directly on mobile devices is accompanied by a number of significant limitations associated with limited computing power, a small amount of RAM, energy constraints, and the peculiarities of the thermal mode of operation. It is separately established that the mobile environment is characterized by high dynamics of available resources. Available computing power, battery charge level, device temperature, activity of background processes, and available memory can change during system operation. This creates additional difficulties for the execution of machine learning models and significantly complicates the use of static optimization methods.

Downloads

Download data is not yet available.

References

InviAI. (n.d.). What is Edge AI? Benefits and challenges. https://inviai.com/uk/sho-take-edge-ai

ProIT. (n.d.). How edge computing is changing computing technology. https://proit.com.ua/news/yak-edge-computing-zminyuye-obchyslyuvalni/

ETSI. (2019). Mobile Edge Computing (MEC); Framework and reference architecture (ETSI GS MEC 003 V2.1.1). https://www.etsi.org/deliver/etsi_gs/MEC/001_099/003/02.01.01_60/gs_MEC003v020101p.pdf

Google. (n.d.). TensorFlow Lite guide: Performance best practices. https://www.tensorflow.org/lite/performance/best_practices

Mao, Z. (2025). Lightweight network architecture – Optimization of deep learning models. Applied and Computational Engineering. https://doi.org/10.54254/2755-2721/2025.AST26509

Liu, L., & Xu, Z. (2025). Optimizing lightweight neural networks for efficient mobile edge computing. Scientific Reports, 15, 22056. https://doi.org/10.1038/s41598-025-04652-7

Howard, A., Zhu, M., Chen, B., et al. (2017). MobileNets: Open-source models for efficient on-device vision applications. Google Research. https://research.google/blog/mobilenets-open-source-models-for-efficient-on-device-vision/

Sandler, M., Howard, A., Zhu, M., et al. (2018). MobileNetV2: Inverted residuals and linear bottlenecks. In IEEE/CVF CVPR. https://doi.org/10.48550/arXiv.1801.04381

Ma, N., Zhang, X., Zheng, H., & Sun, J. (2018). ShuffleNet V2: Practical guidelines for efficient CNN architecture design. In ECCV. https://link.springer.com/chapter/10.1007/978-3-030-01264-9_8

Tan, M., & Le, Q. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In ICML. https://proceedings.mlr.press/v97/tan19a.html

Kuznetsov, O., Frontoni, E., Kuznetsova, Y., Chevardin, V., & Smirnov, O. (2025). Architectural foundations for adaptive security in edge computing systems. In Cybersecurity Defensive Walls in Edge Computing (pp. 21–61). https://www.scopus.com/pages/publications/105026847051

Kuznetsov, O., Atzeni, G., Arnesano, M., Randieri, C., & Smirnov, O. (2025). Secure IoT-based smart wheelchair system: From implementation to security enhancement strategy. In Security and Privacy of Cyber Physical Systems Emerging Trends Technologies and Applications (pp. 225–257). https://www.scopus.com/pages/publications/105014300075

Kuznetsov, O., Smirnov, O., Kuznetsova, T., Shaikhanova, A., & Svatowsky, I. (2025). Privacy-utility trade-offs in IoT networks: A comparative analysis of differential privacy mechanisms for sensor data aggregation. In Security and Privacy of Cyber Physical Systems Emerging Trends Technologies and Applications (pp. 589–622). https://www.scopus.com/pages/publications/105014301897

Smirnov, O., Tkachuk, R., Kozirova, N., Konstantinova, L., Konoplytska-Slobodeniuk, O., Yakymenko, N., & Smirnov, S. (2026). Research on the application of vector databases in generative artificial intelligence. Cybersecurity: Education, Science, Technology, 1(33), 667–682. https://doi.org/10.28925/2663-4023.2026.33.1248

Smirnov, O., Zaritsky, V., Buravchenko, K., Konoplytska-Slobodeniuk, O., Konstantinova, L., Yakymenko, N., & Smirnov, S. (2026). Optimization of face recognition using the CUDA accelerated dlib library. Cybersecurity: Education, Science, Technology, 4(32), 573–582. https://doi.org/10.28925/2663-4023.2026.32.1154

Downloads


Abstract views: 13

Published

2026-09-24

How to Cite

Bohachenko, S., Smirnov, O., Buravchenko, K., Smirnova, T., Konoplitska-Slobodeniuk, O., Yakymenko , N., & Smirnov, S. (2026). LIMITATIONS OF RESOURCES OF MOBILE AND EDGE PLATFORMS WHEN DEVELOPING SOFTWARE FOR IMPLEMENTATION AND OPTIMIZATION OF MACHINE LEARNING MODELS. Electronic Professional Scientific Journal «Cybersecurity: Education, Science, Technique», 2(34), 677–686. https://doi.org/10.28925/2663-4023.2026.34.1360

Issue

Vol. 2 No. 34 (2026): Cybersecurity: Education, Science, Technique

Section

Articles

License

Copyright (c) 2026 Сергій Богаченко , Олексій Смірнов, Костянтин Буравченко , Тетяна Смірнова , Оксана Конопліцька-Слободенюк , Наталія Якименко , Сергій Смірнов

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Most read articles by the same author(s)

1 2 3 > >>