FORECASTING ENERGY DEMAND IN TURKEY USING DIFFERENT METAHEURISTIC METHODS: A COMPARATIVE STUDY

Abstract

Energy demand forecasting plays a crucial role in shaping energy policies, particularly for countries like Turkey that experience rapid industrialization and urbanization. Accurately predicting energy demand helps to ensure energy supply security and to guide strategic investments, especially in transitioning towards renewable energy sources. This study explores the use of modern metaheuristic optimization methods to forecast Turkey's energy demand up to the year 2035, focusing on the effectiveness of various techniques in addressing this complex, multi-dimensional problem. The dataset used spans from 1979 to 2011 and includes economic and demographic indicators such as GDP, population, imports, and exports, which are key drivers of energy demand. Several metaheuristic algorithms, including The African Vultures Optimization Algorithm (AVOA), Grey Wolf Optimizer (GWO), Whale Optimization Algorithm (WOA), and Dynamic Bayesian Optimization (DBO), were applied to this dataset. A comparative analysis of these methods demonstrated that AVOA, GWO, DBO, and other similar approaches yielded the most accurate predictions, with minimum total error rates. The analysis revealed that the AVOA method outperformed other methods in terms of accuracy and computational efficiency by obtaining the lowest total error of 0.2391 and relative error percentage of 0.3565. The study highlights the significant role metaheuristic approaches play in improving the accuracy of energy demand forecasts and informs future policy decisions by identifying critical factors affecting Turkey’s energy consumption patterns. The findings are expected to contribute to more effective long-term energy planning and the development of sustainable energy policies.