Short-term electricity demand forecasting using a hybrid ANFIS-ELM network optimised by an improved parasitism-predation algorithm

One of the most crucial steps in comprehensively planning the proper and efficient use of energy (optimisation of consumption) and appropriate management of resources to fulfil the power demand is accurately forecasting the energy demand (optimisation of production). This study presents the modelling of an ideal technique for shortterm power demand prediction as a novel hybrid approach comprising two distinct methods, namely the Elman neural network (ELM) and adaptive network-based fuzzy inference system (ANFIS). The hybrid approach outperforms the conventional individual methods because it can eliminate the drawbacks of the individual methods while retaining their advantages, particularly considering that ELM and ANFIS can handle non-linear data. The weight coefficients of the procedures in the proposed hybrid approach were determined using a novel, enhanced bioinspired algorithm, called the improved parasitism-predation algorithm, to achieve better accuracy. The simulation results demonstrated the superiority of the proposed approach over other state-of-the-art approaches including the independent ELM and ANFIS.

Automated summary

This study presents a novel hybrid approach for short-term power demand prediction, using Elman neural network (ELM) and adaptive network-based fuzzy inference system (ANFIS). The hybrid approach outperforms conventional individual methods by eliminating their drawbacks while retaining their advantages, especially in handling non-linear data. The proposed approach is determined using an enhanced bioinspired algorithm, called the improved parasitism-predation algorithm, achieving better accuracy. Simulation results demonstrate the superiority of this approach over other state-of-the-art approaches, including independent ELM and ANFIS.