<p>Smart city electricity forecasting&#xa0;is pivotal in&#xa0;enabling sustainable urban development, efficient energy distribution, and real-time decision-making for infrastructure resilience. This study proposes a hybrid time series forecasting model that integrates&#xa0;the Auto-Regressive Integrated Moving Average (ARIMA) method with the bio-inspired Greylag Goose Optimization (GGO) algorithm to enhance prediction&#xa0;accuracy for electricity demand in innovative urban environments. The proposed GGO-ARIMA model is&#xa0;applied to an hourly electricity load dataset&#xa0;enriched with exogenous features such&#xa0;as weather, public holidays, and academic schedules—factors influencing urban energy consumption.&#xa0;Experimental evaluations reveal that GGO-ARIMA significantly outperforms&#xa0;conventional ARIMA and other optimization-enhanced models, achieving a Mean&#xa0;Squared Error (MSE) of 0.002135 and <i>R</i><sup>2</sup> of 0.99995, reflecting substantial&#xa0;reductions in forecast error. This hybrid approach&#xa0;offers a scalable and interpretable solution&#xa0;for smart city energy management systems. The findings underscore the potential of optimization-enhanced time&#xa0;series models for advancing intelligent urban&#xa0;infrastructures.</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

Smart City Electricity Load Forecasting Using Greylag Goose Optimization-Enhanced Time Series Analysis

  • El-Sayed M. El-Kenawy,
  • Abdelhameed Ibrahim,
  • Amel Ali Alhussan,
  • Doaa Sami Khafaga,
  • Ayman E. M. Ahmed,
  • Marwa M. Eid

摘要

Smart city electricity forecasting is pivotal in enabling sustainable urban development, efficient energy distribution, and real-time decision-making for infrastructure resilience. This study proposes a hybrid time series forecasting model that integrates the Auto-Regressive Integrated Moving Average (ARIMA) method with the bio-inspired Greylag Goose Optimization (GGO) algorithm to enhance prediction accuracy for electricity demand in innovative urban environments. The proposed GGO-ARIMA model is applied to an hourly electricity load dataset enriched with exogenous features such as weather, public holidays, and academic schedules—factors influencing urban energy consumption. Experimental evaluations reveal that GGO-ARIMA significantly outperforms conventional ARIMA and other optimization-enhanced models, achieving a Mean Squared Error (MSE) of 0.002135 and R2 of 0.99995, reflecting substantial reductions in forecast error. This hybrid approach offers a scalable and interpretable solution for smart city energy management systems. The findings underscore the potential of optimization-enhanced time series models for advancing intelligent urban infrastructures.