A Novel Fractional ARIMA Model with Genetic Algorithm and Its Applications in Forecasting the Electricity Consumption Demand

Electric system planning requires accurate electricity consumption forecasting. Accurate electricity consumption forecasting is required for policymakers to develop electricity distribution policies. However, insufficient data, which is commonly highly nonlinear, cannot provide sufficient data to identify satisfying forecasting accuracy. To solve these problems, several researchers used a grey model. Fractional cumulative generation operation (FAGO) is a relatively new and common method for improving the accuracy of grey models. In order to forecast electricity consumption, this study proposed a new hybrid FARIMA model that combines the FAGO and the autoregressive integrated moving average (ARIMA). Additionally, the fractional order is optimized through the use of the Genetic Algorithm (GA). Two distinct examples are used to compare the proposed model's efficacy to traditional ARIMA and two well-known grey models. The FARIMA model outperforms the other three models in all case studies, demonstrating that it can be used as a precise and promising approach for forecasting in the short term with small datasets.