International Journal of Engineering and Manufacturing (IJEM)
IJEM Vol. 16, No. 4, 8 Aug. 2026
Cover page and Table of Contents: PDF (size: 1079KB)
Software for Simulating Neural Network Workload Impact on Mobile Devices
PDF (1079KB), PP.287-301
Views: 0 Downloads: 0
Author(s)
Kirill Smelyakov
1
Oleksandr Dolhanenko
1
Oleksiy Lanovyy
1
Victoria Vysotska
2,3
Dmytro Uhryn
4,*
Index Terms
Android, battery drain, machine learning, mobile computing, neural networks, on-device training, performance evaluation, power management, reproducibility, resource allocation, task scheduling, thermal modelling, trace-driven simulation
Abstract
Mobile processors are now capable of running complex machine learning tasks locally, yet our mobile operating systems often hold them back. To preserve battery life, heavy workloads are typically restricted to charging or idle periods, severely limiting time-sensitive applications such as real-time health monitoring or federated learning. While we need more innovative scheduling algorithms to overcome this, validating them on physical hardware is extremely difficult. Factors like thermal throttling, background kernel activity, and battery degradation create an unpredictable environment where no two tests are ever quite the same. To solve this reproducibility challenge, we introduce a new simulation framework. Instead of relying on inconsistent test runs on physical devices, our system records a device's natural baseline activity and mathematically superimposes the resource footprint of a heavy task. It allows us to model CPU saturation, energy loss, memory pressure, and thermal dynamics in a controlled software environment. When compared against ground-truth recordings from Samsung Galaxy S10e and Fold 5 devices, the simulator achieved correlation scores exceeding 0.90 for computing and thermal metrics. By isolating the workload's impact from environmental noise, this platform provides a scalable method for benchmarking neural network schedulers without the logistical bottlenecks associated with continuous hardware testing.
Cite This Paper
Kirill Smelyakov, Oleksandr Dolhanenko, Oleksiy Lanovyy, Victoria Vysotska, Dmytro Uhryn, "Software for Simulating Neural Network Workload Impact on Mobile Devices", International Journal of Engineering and Manufacturing (IJEM), Vol.16, No.4, pp.287-301, 2026. DOI:10.5815/ijem.2026.04.20
Reference
[1]T. Tan and G. Cao. Efficient execution of deep neural networks on mobile devices with NPU. In Proceedings of the 20th International Conference on Information Processing in Sensor Networks (IPSN 2021), pages 283–298, Nashville, USA, 2021.
[2]P. Nawrocki, B. Śnieżyński, J. Kołodziej, and P. Szynkiewicz. Adaptive context-aware energy optimisation for services on mobile devices using machine learning considering security aspects. In 2020 IEEE/ACM International Symposium on Cluster, Cloud and Internet Computing (CCGRID), pages 708–717, Melbourne, Australia, 2020.
[3]C. Luo, J. Gonçalves, E. Velloso, and V. Kostakos. A survey of context simulation for testing mobile context-aware applications. ACM Computing Surveys, 53(1):21, 2020.
[4]K. Lee, S.-R. Ohk, S.-G. Lim, and Y.-J. Kim. Phase-based accurate power modeling for mobile application processors. Electronics, 10(10):1197, 2021.
[5]K. Huang, B. Yang, and W. Gao. ElasticTrainer: Speeding up on-device training with runtime elastic tensor selection. In Proceedings of the 21st International Conference on Mobile Systems, Applications and Services (MobiSys 2023), pages 56–69, 2023.
[6]Q. Zhang, Z. Zhu, A. Zhou, Q. Sun, S. Dustdar, and S. Wang. Energy-efficient federated training on mobile devices. IEEE Network, pages 1–7, 2023.
[7]S. Zhu, T. Voigt, F. Rahimian, and J. Ko. On-device training: A first overview on existing systems. ACM Transactions on Sensor Networks, 20(6):118, 2024.
[8]D. Cai, Q. Wang, Y. Liu, Y. Liu, S. Wang, and M. Xu. Towards ubiquitous learning: A first measurement of on-device training performance. In Proceedings of the 5th International Workshop on Embedded and Mobile Deep Learning (EMDL 2021), pages 31–36, 2021.
[9]J. Liu, J. Liu, W. Du, and D. Li. Performance analysis and characterisation of training deep learning models on mobile devices. arXiv preprint arXiv:1906.04278, 2019.
[10]H.-H. Sung, J.-A. Chen, W. Niu, J. Guan, B. Ren, and X. Shen. Decentralised application-level adaptive scheduling for multi-instance DNNs on open mobile devices. In Proceedings of the USENIX Annual Technical Conference (USENIX ATC 2023), pages 865–877, Boston, USA, 2023.
[11]T. Tan and G. Cao. Thermal-aware scheduling for deep learning on mobile devices with NPU. IEEE Transactions on Mobile Computing, 23(12):10706–10719, 2024.
[12]F. Jiang, Y. Xia, L. Yan, W. Liu, X. Zhang, and H. Li. Battery-aware workflow scheduling for portable heterogeneous computing. IEEE Transactions on Sustainable Computing, pages 1–18, 2024.
[13]V. Ravihansa, S. Wang, and M. Liyanage. Energy-Aware Federated Learning: Solutions and Challenges. 2025 IEEE 4th International Conference on Industrial Electronics and Applications (IESES), pages 1–8, 2025.
[14]K. Liu, X. Zhou, and L. Li. m2LLM: A multi-dimensional optimisation framework for LLM inference on mobile devices. IEEE Transactions on Parallel & Distributed Systems, 36(10):2014–2029, 2025
[15]X. Tu, A. Mallik, H. Wang, and J. Xie. AIEnergy: An energy benchmark for AI-empowered mobile and IoT devices. ITU Journal on Future and Evolving Technologies, 6(2):183–197, 2025
[16]M. Hirsch, C. Mateos, J. M. Rodríguez, and A. Zunino. DewSim: A trace-driven toolkit for simulating mobile device clusters in dew computing environments. Software: Practice and Experience, 50(5):688–718, 2020.
[17]M. Hirsch, C. Mateos, A. Zunino, and J. Toloza. A platform for automating battery-driven batch benchmarking and profiling of Android-based mobile devices. Simulation Modelling Practice and Theory, 109:102266, 2021.
[18]Y. Wu et al. MOC: Multi-objective mobile CPU–GPU co-optimisation for power-efficient DNN inference. In IEEE/ACM International Conference on Computer-Aided Design (ICCAD 2023), pages 1–10, San Francisco, USA, 2023.
[19]A. Jaiswal, Z. Amjad, and R. Nair, "Diminishing Returns and Deep Learning for Adaptive CPU Resource Allocation of Containers," 2020 IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS), Washington, DC, USA, 2020, pp. 11-20.
[20]J. Fang, Y. Chen, and S. Lu. A scheduling strategy for reduced power consumption in mobile edge computing. In IEEE INFOCOM 2020 Workshops, pages 1190–1195, Toronto, Canada, 2020.
[21]M. Zhao et al. Energy-aware task offloading and resource allocation for time-sensitive services in mobile edge computing systems. IEEE Transactions on Vehicular Technology, 70(10):10925–10940, October 2021.
[22]Muzychuk, O., Vysotska, V., Vasylenko, V., Tsuranov, M., Zemlianko, H., Khavina, I., & Gnusov, Y. (2025). Smart indoor evacuation using real-time beacons. Ceur Workshop procedings, Vol-4126. https://ceur-ws.org/Vol-4126/paper9.pdf
[23]Vysotska, V., Smelyakov, K., Sharonova, N., Vakulik, E., Filipov, O., & Kotelnykov, R. (2024). Fast Color Images Clustering for Real-Time Computer Vision and AI System. In Ceur Workshop procedings (1) (pp. 161-177). https://ceur-ws.org/Vol-3664/paper12.pdf
[24]Vladov, S., Yakovliev, R., Yurko, A., Vysotska, V., Lytvyn, V., & Romanchuk, R. (2025, April). Hybrid Networks Training Method Based on Distributed and Parallel Computing Technology. In Lecture Notes in Data Engineering, Computational Intelligence, and Decision-Making, Volume 2: 2024 International Scientific Conference" Intelligent Systems of Decision-Making and Problems of Computational Intelligence”, Proceedings (Vol. 2, p. 3). Springer Nature.