Scholarlink

<p>Energy conservation is a global issue and one of the current research themes of the last decade. Population growth, indoor environmental quality, and climate change are dramatically increasing the need for sustainable technologies and solutions to save energy in residential buildings. Recently, Internet of things (IoT) applications have been developed in smart homes, smart cities, smart hospitals, and other smart environments. Sustainable technology goals in residential buildings include maximizing thermal comfort and minimizing energy consumption. Residential building challenges and problems can be solved using consumer behavioral models and integration into housing problem solutions. This article proposes an IoT task management mechanism based on predictive optimization to minimize energy consumption in smart residential buildings. A hybrid forecasting system for power consumption is suggested. The system combines time series forecasting techniques (an autoregressive integrated moving average (ARIMA)), (a seasonal autoregressive integrated moving average (SARIMA)), and deep learning techniques like Deep Extreme Learning Machine (DELM). Experiments confirm that a CNN-LSTM neural network extracts complex features of energy consumption. CNN layers extract features from multiple variables that affect energy consumption, and LSTM layers are well-suited for modeling irregular trend temporal information in time series components. The proposed CNN-LSTM method achieves near-perfect prediction performance of power consumption.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Network-Aware Energy Management in IoT Systems: Benchmarking Time Series and Deep Learning Methods for Network Resource Optimization

  • 期刊论文

  • 发表日期：

    2026年7月27日

查看全文
  • Amira Zrelli,
  • Hacen Khlaifi,
  • Ghada Goui,
  • Tahar Ezzedine

摘要

Energy conservation is a global issue and one of the current research themes of the last decade. Population growth, indoor environmental quality, and climate change are dramatically increasing the need for sustainable technologies and solutions to save energy in residential buildings. Recently, Internet of things (IoT) applications have been developed in smart homes, smart cities, smart hospitals, and other smart environments. Sustainable technology goals in residential buildings include maximizing thermal comfort and minimizing energy consumption. Residential building challenges and problems can be solved using consumer behavioral models and integration into housing problem solutions. This article proposes an IoT task management mechanism based on predictive optimization to minimize energy consumption in smart residential buildings. A hybrid forecasting system for power consumption is suggested. The system combines time series forecasting techniques (an autoregressive integrated moving average (ARIMA)), (a seasonal autoregressive integrated moving average (SARIMA)), and deep learning techniques like Deep Extreme Learning Machine (DELM). Experiments confirm that a CNN-LSTM neural network extracts complex features of energy consumption. CNN layers extract features from multiple variables that affect energy consumption, and LSTM layers are well-suited for modeling irregular trend temporal information in time series components. The proposed CNN-LSTM method achieves near-perfect prediction performance of power consumption.