kaz29Вестник Казахстанско-Британского технического университетаHerald of the Kazakh-British Technical University1998-66882959-8109Казахстанско-Британский Технический Университет10.55452/1998-6688-2026-23-2-133-149kaz29-2892Research ArticleКОМПЬЮТЕРНЫЕ НАУКИCOMPUTER SCIENCEПРОГНОЗИРОВАНИЕ ЭНЕРГИИ НА ОСНОВЕ ИИ И УЛУЧШЕННОЕ УПРАВЛЕНИЕ СПРОСОМ В УМНЫХ ДОМАХAI-BASED ENERGY FORECASTING AND IMPROVED DEMAND MANAGEMENT IN SMART HOMEShttps://orcid.org/0000-0003-0764-8574ТохметовА.TokhmetovA.

к. ф.-м. н., ассоциированный профессор.

Астана

Cand. Phys.-Math. Sc., Associate Professor.

Astana

tokhmetov_at_2@enu.kzhttps://orcid.org/0000-0002-3627-3321СерикбаеваС.SerikbayevaS.

PhD, ассоциированный профессор.

Астана

PhD, Associate Professor.

Astana

inf_8585@mail.ruhttps://orcid.org/0000-0002-6811-2303ТанченкоЛ.TanchenkoL.

Магистр.

Астана

MSc.

Astana

ltanchenko@mail.ruhttps://orcid.org/0009-0000-2121-089XКенесбайМ.KenesbayM.

Магистрант.

Астана

Master’s student.

Astana

mikam4965@gmail.comЕвразийский национальный университет им. Л.Н. ГумилеваКазахстанL.N. Gumilyov Eurasian National UniversityKazakhstan202627062026232133149Copyright © Тохметов А., Серикбаева С., Танченко Л., Кенесбай М., 20262026Тохметов А., Серикбаева С., Танченко Л., Кенесбай М.Tokhmetov A., Serikbayeva S., Tanchenko L., Kenesbay M.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://vestnik.kbtu.edu.kz/jour/article/view/2892

В данной статье представлена комплексная многоступенчатая система, разработанная для повышения точности прогнозов энергетической нагрузки и оценки эффективности как моделей прогнозирования, так и стратегий реагирования на спрос (DR). Используя набор данных REFIT, был проведен сравнительный анализ иерархии моделей прогнозирования, включая линейную регрессию, случайный лес, SVR, k-NN, LSTM и гибридный кодер-декодер с механизмом внимания. Результаты исследования показали, что разработанная гибридная модель кодера-декодера с механизмом внимания достигла наилучшей точности (R² = 0.91, MAPE = 2.39%), продемонстрировав отличную способность улавливать сложные временные закономерности в данных. Тщательное многоступенчатое тестирование подтвердило стабильность и высокую обобщаемость этой модели глубокого обучения. Высокоточный прогноз был встроен в модель на основе смешанного целочисленного линейного программирования (MILP) для оптимизации системы управления энергопотреблением дома (HEMS). Результаты показали, что эта комплексная структура позволила значительно сократить затраты на электроэнергию на 22.5% и снизить пиковую нагрузку на 31.8% за счет оптимального планирования работы бытовых приборов. Эта работа демонстрирует, как эффективно сочетать передовое прогнозирование на основе искусственного интеллекта (ИИ) с формальной оптимизацией энергопотребления в единой, комплексной системе. Этот метод не только позволяет более точно прогнозировать потребление, особенно в часы пик, но также демонстрирует, что ИИ может значительно повысить гибкость энергетических сетей и энергоэффективность умных домов.

This paper presents a comprehensive multi-stage system designed to improve the accuracy of energy load forecasts and evaluate the effectiveness of both forecast models and demand response (DR) strategies. Using the REFIT dataset, a comparative analysis of a hierarchy of forecast models was conducted, including linear regression, random forest, SVR, k-NN, LSTM, and a hybrid encoder-decoder with an attention mechanism. The results of the study indicated that the developed hybrid encoder-decoder model with an attention mechanism achieved the best accuracy (R² = 0.91, MAPE = 2.39%), demonstrating excellent ability to capture complex temporal patterns in the data. Rigorous multi-stage testing confirmed the stability and high generalizability of this deep learning model. The highly accurate forecast was incorporated into a mixed integer linear programming (MILP)-based model for home energy management system (HEMS) optimization. The results indicated that this complex framework significantly reduced energy costs by 28.7% and reduced peak load by 37.1% through optimal appliance scheduling. This work demonstrates how to effectively combine state-of-the-art artificial intelligence (AI)-based forecasting with formal energy optimization in a single, comprehensive system. This method not only allows for more accurate consumption forecasting, especially during peak hours, but also demonstrates that AI can significantly improve the flexibility of energy networks and the energy efficiency of smart homes.

прогнозирование энергииреагирование на спросумные домамашинное обучениеглубокое обучениеLSTMгибридная модельenergy forecastingdemand responsesmart homesmachine learningdeep learningLSTMhybrid modelReferencesArastehfar, S., Matinkia, M., Jabbarpour, M. Short-term residential load forecasting using Graph Convolutional Recurrent Neural Networks. Engineering Applications of Artificial Intelligence, 116 (1), 105358 (2022). https://doi.org/10.1016/j.engappai.2022.105358.Arastehfar, S., Matinkia, M., Jabbarpour, M. Short-term residential load forecasting using Graph Convolutional Recurrent Neural Networks. Engineering Applications of Artificial Intelligence, 116 (1), 105358 (2022). https://doi.org/10.1016/j.engappai.2022.105358.Gonzalez, R., Ahmed, S., & Alamaniotis, M. Implementing Very-Short-Term Forecasting of Residential Load Demand Using a Deep Neural Network Architecture. Energies, 16 (9), 3636 (2023). https://doi.org/10.3390/en16093636.Gonzalez, R., Ahmed, S., & Alamaniotis, M. Implementing Very-Short-Term Forecasting of Residential Load Demand Using a Deep Neural Network Architecture. Energies, 16 (9), 3636 (2023). https://doi.org/10.3390/en16093636.Chatuanramtharnghaka, B., Deb, S., & Singh, K. Short-Term Load Forecasting for IEEE 33 Bus Test System using SARIMAX. IEEE 2nd International Conference on Industrial Electronics: Developments & Applications (ICIDeA), Imphal, India, 275–280 (2023). https://doi.org/10.1109/ICIDeA59866.2023.10295066.Chatuanramtharnghaka, B., Deb, S., & Singh, K. Short-Term Load Forecasting for IEEE 33 Bus Test System using SARIMAX. IEEE 2nd International Conference on Industrial Electronics: Developments & Applications (ICIDeA), Imphal, India, 275–280 (2023). https://doi.org/10.1109/ICIDeA59866.2023.10295066.Lee, G-C. A Regression-Based Method for Monthly Electric Load Forecasting in South Korea. Energies, 17 (23), 5860-5875 (2024). https://doi.org/10.3390/en17235860.Lee, G-C. A Regression-Based Method for Monthly Electric Load Forecasting in South Korea. Energies, 17 (23), 5860-5875 (2024). https://doi.org/10.3390/en17235860.Al-Turjman, F., & Malekloo, A. Machine learning for energy prediction in smart homes: A survey. Sustainable Cities and Society, 101, 104457 (2024). https://doi.org/10.1016/j.scs.2023.104457.Al-Turjman, F., & Malekloo, A. Machine learning for energy prediction in smart homes: A survey. Sustainable Cities and Society, 101, 104457 (2024). https://doi.org/10.1016/j.scs.2023.104457.Ji, Y., Zhu, Y., Lu, S., Yang, L., Liew, A.W.-C.: Wtc-ipst: A deep learning framework for short-term electric load forecasting with multi-scale feature extraction. Knowledge-Based Systems, 24, 113907 (2025). https://doi.org/10.1016/j.knosys.2025.113907.Ji, Y., Zhu, Y., Lu, S., Yang, L., Liew, A.W.-C.: Wtc-ipst: A deep learning framework for short-term electric load forecasting with multi-scale feature extraction. Knowledge-Based Systems, 24, 113907 (2025). https://doi.org/10.1016/j.knosys.2025.113907.Faria, P., & Vale, Z. Demand Response in Smart Grids. Energies, 16 (2), 863 (2023). https://doi.org/10.3390/en16020863.Faria, P., & Vale, Z. Demand Response in Smart Grids. Energies, 16 (2), 863 (2023). https://doi.org/10.3390/en16020863.Wang, Y., Zhang, N., Zhuo, Z., Kang, C., Kirschen, D. Mixed-Integer Linear Programming-Based Optimal Configuration Planning for Energy Hub: Starting from Scratch. Applied Energy, 210 (2), 1141–1150 (2018). https://doi.org/10.1016/j.apenergy.2017.08.114.Wang, Y., Zhang, N., Zhuo, Z., Kang, C., Kirschen, D. Mixed-Integer Linear Programming-Based Optimal Configuration Planning for Energy Hub: Starting from Scratch. Applied Energy, 210 (2), 1141–1150 (2018). https://doi.org/10.1016/j.apenergy.2017.08.114.Chandrasekaran, R., Paramasivan, S.K. Advances in deep learning techniques for short-term energy load forecasting applications: A review. Archives of Computational Methods in Engineering, 32 (2), 663-692 (2025). https://doi.org/10.1007/s11831-024-10155-x.Chandrasekaran, R., Paramasivan, S.K. Advances in deep learning techniques for short-term energy load forecasting applications: A review. Archives of Computational Methods in Engineering, 32 (2), 663-692 (2025). https://doi.org/10.1007/s11831-024-10155-x.Chen, Y., Liu, H., & Wu, Y. Integrating renewable energy forecasting with smart home demand response. Sustainable Energy Technologies and Assessments, 65, 102918 (2025). https://doi.org/10.1016/j.seta.2025.102918.Chen, Y., Liu, H., & Wu, Y. Integrating renewable energy forecasting with smart home demand response. Sustainable Energy Technologies and Assessments, 65, 102918 (2025). https://doi.org/10.1016/j.seta.2025.102918.Branco, N.W., Cavalca, M.S., Stefenon, S.F., & Leithardt, V.R. Wavelet LSTM for Fault Forecasting in Electrical Power Grids. Sensors, 22 (21), 8323 (2022). https://doi.org/10.3390/s22218323.Branco, N.W., Cavalca, M.S., Stefenon, S.F., & Leithardt, V.R. Wavelet LSTM for Fault Forecasting in Electrical Power Grids. Sensors, 22 (21), 8323 (2022). https://doi.org/10.3390/s22218323.Ning, Y., Kazemi, H., & Tahmasebi, P. A comparative machine learning study for time series oil production forecasting: ARIMA, LSTM, and Prophet. Computers and Geosciences, 164, 105126 (2022). https://doi.org/10.1016/j.cageo.2022.105126.Ning, Y., Kazemi, H., & Tahmasebi, P. A comparative machine learning study for time series oil production forecasting: ARIMA, LSTM, and Prophet. Computers and Geosciences, 164, 105126 (2022). https://doi.org/10.1016/j.cageo.2022.105126.Sun, Y., Zhang, H., & Li, Y. Deep learning approaches for household load forecasting: A comparative analysis. Energy and Buildings, 254, 111608 (2022). https://doi.org/10.1016/j.enbuild.2021.111608.Sun, Y., Zhang, H., & Li, Y. Deep learning approaches for household load forecasting: A comparative analysis. Energy and Buildings, 254, 111608 (2022). https://doi.org/10.1016/j.enbuild.2021.111608.Ghadertootoonchi, A., Moeini-Aghtaie, M., Davoudi, M. AHybrid Linear Programming-Reinforcement Learning Method for Optimal Energy Hub Management. IEEE Transactions on Smart Grid, 14 (1), 157–166 (2023). https://doi.org/10.1109/TSG.2022.3197458.Ghadertootoonchi, A., Moeini-Aghtaie, M., Davoudi, M. AHybrid Linear Programming-Reinforcement Learning Method for Optimal Energy Hub Management. IEEE Transactions on Smart Grid, 14 (1), 157–166 (2023). https://doi.org/10.1109/TSG.2022.3197458.Ma, P., Cui, S., Chen, M., Zhou, S., Wang, K. Review of family-level short-term load forecasting and its application in household energy management system. Energies, 16 (15), 5809–5825 (2023). https://doi.org/10.3390/en16155809.Ma, P., Cui, S., Chen, M., Zhou, S., Wang, K. Review of family-level short-term load forecasting and its application in household energy management system. Energies, 16 (15), 5809–5825 (2023). https://doi.org/10.3390/en16155809.Du, S., Li, T., Yang, Y., & Horng, S. Multivariate time series forecasting via attention-based encoderdecoder framework. Neurocomputing, 388, 269–279 (2020). https://doi.org/10.1016/j.neucom.2019.12.118.Du, S., Li, T., Yang, Y., & Horng, S. Multivariate time series forecasting via attention-based encoderdecoder framework. Neurocomputing, 388, 269–279 (2020). https://doi.org/10.1016/j.neucom.2019.12.118.Perçuku, A., Minkovska, D., Hinov, N. Enhancing Electricity Load Forecasting with Machine Learning and Deep Learning. Technologies, 13 (2), 70–90 (2025). https://doi.org/10.3390/technologies13020059.Perçuku, A., Minkovska, D., Hinov, N. Enhancing Electricity Load Forecasting with Machine Learning and Deep Learning. Technologies, 13 (2), 70–90 (2025). https://doi.org/10.3390/technologies13020059.Santoro, D., Ciano, T. & Ferrara, M. A comparison between machine and deep learning models on high stationarity data. Scientific Reports, 14 (1), 19409–1420 (2024). https://doi.org/10.1038/s41598-024-70341-6.Santoro, D., Ciano, T. & Ferrara, M. A comparison between machine and deep learning models on high stationarity data. Scientific Reports, 14 (1), 19409–1420 (2024). https://doi.org/10.1038/s41598-024-70341-6.Bodenschatz, N., Eider M., & Berl, A. Mixed-Integer-Linear-Programming Model for the Charging Scheduling of Electric Vehicle Fleets. 2020 10th International Conference on Advanced Computer Information Technologies (ACIT), Deggendorf, Germany, 741–746 (2020). https://doi.org/10.1109/ACIT49673.2020.9208875.Bodenschatz, N., Eider M., & Berl, A. Mixed-Integer-Linear-Programming Model for the Charging Scheduling of Electric Vehicle Fleets. 2020 10th International Conference on Advanced Computer Information Technologies (ACIT), Deggendorf, Germany, 741–746 (2020). https://doi.org/10.1109/ACIT49673.2020.9208875.Dukpa, A., & Butrylo, B. MILP-Based Profit Maximization of Electric Vehicle Charging Station Based on Solar and EV Arrival Forecasts. Energies, 15 (15), 5760 (2022). https://doi.org/10.3390/en15155760.Dukpa, A., & Butrylo, B. MILP-Based Profit Maximization of Electric Vehicle Charging Station Based on Solar and EV Arrival Forecasts. Energies, 15 (15), 5760 (2022). https://doi.org/10.3390/en15155760.Murray, D., Stankovic, L. & Stankovic, V. An electrical load measurements dataset of United Kingdom households from a two-year longitudinal study. Scientific Data, 4, 160122–160130 (2017). https://doi.org/10.1038/sdata.2016.122.Murray, D., Stankovic, L. & Stankovic, V. An electrical load measurements dataset of United Kingdom households from a two-year longitudinal study. Scientific Data, 4, 160122–160130 (2017). https://doi.org/10.1038/sdata.2016.122.Willmott, C.J. & Matsuura, K. Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance. Climate Research, 30 (1), 79–82 (2005). https://doi.org/10.3354/cr030079.Willmott, C.J. & Matsuura, K. Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance. Climate Research, 30 (1), 79–82 (2005). https://doi.org/10.3354/cr030079.Chai, T. & Draxler, R.R. Root mean square error (RMSE) or mean absolute error (MAE)? Arguments against avoiding RMSE in the literature. Geoscientific Model Development, 7, 1247–1250 (2014). https://doi.org/10.5194/gmd-7-1247-2014.Chai, T. & Draxler, R.R. Root mean square error (RMSE) or mean absolute error (MAE)? Arguments against avoiding RMSE in the literature. Geoscientific Model Development, 7, 1247–1250 (2014). https://doi.org/10.5194/gmd-7-1247-2014.Kvålseth, T.O. Cautionary note about R². The American Statistician, 39 (4), 279–285 (1985). https://doi.org/10.1080/00031305.1985.10479448.Kvålseth, T.O. Cautionary note about R². The American Statistician, 39 (4), 279–285 (1985). https://doi.org/10.1080/00031305.1985.10479448.Tofallis, C. A better measure of relative prediction accuracy for model selection and model estimation. Journal of the Operational Research Society, 66 (8), 1352–1362 (2015). https://doi.org/10.1057/jors.2014.103.Tofallis, C. A better measure of relative prediction accuracy for model selection and model estimation. Journal of the Operational Research Society, 66 (8), 1352–1362 (2015). https://doi.org/10.1057/jors.2014.103.

The authors declare that there are no conflicts of interest present.