PREDICTION OF ELECTRICITY CONSUMPTION USING THREE META-HEURISTIC ALGORITHMS

Energy production and consumption play an important role in the domestic and international strategic decisions globally. Monitoring the electric energy consumption is essential for the short- and long-term of sustainable development planned in different countries. One of the advanced methods and/or algorithms applied in this prediction is the meta-heuristic algorithm. The meta-heuristic algorithms can minimize the errors and standard deviations in the data processing. Statistically, there are numerous methods applicable in the uncertainty analysis and in realizing the errors in the datasets, if any. In this article, the Mean Absolute Percentage Error (MAPE) is used in the error’s minimization within the relevant algorithms, and the used dataset is actually relating to the past fifty years, say from 1972 to 2021. For this purpose, the three algorithms such as the Imputation–Regularized Optimization (IRO), Colliding Bodies Optimization (CBO), and Enhanced Colliding Bodies Optimization (ECBO) have been used. Each one of the algorithms has been implemented for the two linear and exponential models. Among this combination of the six models, the linear model of the ECBO meta-heuristic algorithm has yielded the least error. The magnitude of this error is about 3.7%. The predicted energy consumption with the winning model planned for the year 2030 is about 459 terawatt-hours. The important socio-economical parameters are used in predicting the energy consumption, where these parameters include the electricity price, Gross Domestic Product (GDP), previous year's consumption, and also the population. Application of the meta-heuristic algorithms could help the electricity generation industries to calculate the energy consumption of the approaching years with the least error. Researchers should use various algorithms to minimize this error and make the more realistic prediction.