A Comparative of Neural Network with Metaheuristics for Electricity Consumption Forecast Modelling

This paper attempted to study the efficiency of Artificial Neural Network (ANN) with metaheuristic algorithms for electricity consumption modeling in Thailand. The objective was to compare the study between ANN with Backpropagation algorithm (ANN-BP) and ANN combined with different metaheuristic algorithms: ANN with Harmony Search (ANN-HS), ANN with Artificial Bee Colony (ANN-ABC), ANN with Teaching-Learning-Based Optimization (ANN-TLBO) and ANN with Jaya Algorithms (ANN-JA) models. The models selected Population, Gross Domestic Product (GDP), Imports of goods and services, Exports of goods and services as inputs. The experiment results showed that the ANN-TLBO model had optimal efficiency, while ANN-JA was one of the competitive metaheuristic algorithms that could be implemented to modelling for use in future studies.