Yu Dongxian, Wu Di, Liao Chongyang, Cao Zaihui, Pouramini Somayeh
College of Modern Information technology, Henan Polytechnic, Zhengzhou, 450046, Henan, China.
JiLin Information & Telecommunication Company, State Grid Jilin Electric Power Corporation Ltd, Changchun, Jilin, 130000, China.
Sci Rep. 2025 Jul 9;15(1):24785. doi: 10.1038/s41598-025-10557-2.
The research paper introduces a novel technique for forecasting electricity usage by utilizing the Developed human evolutionary optimization (DHEO) algorithm and the Xception Neural Network (Xception-NN) model. The Xception-NN model, which is a modified deep learning framework, processes time-series data and incorporates various factors such as weather conditions, demographic insights, and economic indicators. By refining the model's parameters, the DHEO algorithm, inspired by human evolutionary principles, enables a more accurate capture of intricate dependencies and patterns in electricity consumption data. This approach provides energy companies and utilities with a means to enhance their predictions, optimize energy production, and effectively anticipate future demand. Additionally, the study investigates electricity consumption under two scenarios: Base Line (BL) and Energy Conservation (EC), with a focus on the volume of electricity consumed across different sectors. The EC scenario leads to a notable 6.54% reduction in electricity consumption, with the industry sector experiencing the most significant decline.
该研究论文介绍了一种通过利用改进的人类进化优化（DHEO）算法和Xception神经网络（Xception-NN）模型来预测电力使用情况的新技术。Xception-NN模型是一种经过改进的深度学习框架，用于处理时间序列数据，并纳入各种因素，如天气状况、人口统计见解和经济指标。受人类进化原理启发的DHEO算法通过优化模型参数，能够更准确地捕捉电力消耗数据中的复杂依赖关系和模式。这种方法为能源公司和公用事业公司提供了一种增强预测、优化能源生产以及有效预测未来需求的手段。此外，该研究调查了两种情景下的电力消耗情况：基线（BL）情景和节能（EC）情景，重点关注不同部门的电力消耗总量。在EC情景下，电力消耗显著降低了6.54%，其中工业部门的降幅最为显著。
