Prediction of the Electricity Demand in the Market: An Application of Optimization and Machine Learning

Authors

• Ahmed Majed Althahabi ¹
• Hassan Mohammed Abed ²
• Raed Khalid ³
• Abrar Ryadh ⁴
• Ali Al Mansor ⁵
• Kadhum Al-Majdi ⁶
• Adil Abbas Alwan ²

1. Al-Manara College for Medical Sciences, Maysan, Iraq.
2. Mazaya University College, Iraq
3. Department of Medical Instruments Engineering Techniques, Al-Farahidi University, Baghdad, Iraq
4. Medical Laboratory Techniques Department, Al-Mustaqbal University College, 51001 Hillah, Babylon, Iraq
5. Department of Optical Techniques, Al-Zahrawi University College, Karbala, Iraq
6. Department of Biomedical Engineering, Ashur University College, Baghdad, Iraq

Abstract

In this study, the combination of Gray Wolf Optimization and Artificial neural networks (GWO-ANN) algorithm was applied to predict the long-term electricity demand in Iraq, considering the nonlinear trend and uncertainties in the variables affecting it. The results indicate that the population and gross domestic product are significant explanatory variables for long-term energy demand, consistent with previous studies. Compared to other intelligent methods, the GWO-ANN algorithm requires less data for modeling and optimally designs the ANN structure. The modeling and forecasting model outperform the ANN in simulating and predicting the long-term energy demand. Based on the most likely scenario, the predicted electricity demand in Iraq will reach approximately 415 GWh. Electricity is a critical factor in the development of societies and is utilized in various economic sectors.

Keywords

• Modified PWM method
• Electricity demand
• Gray Wolf Optimization
• Artificial Neural Networks
• Predictive modeling