US Natural Gas Consumption Analysis via a Smart Time Series Approach Based on Multilayer Perceptron ANN Tuned by Meta-heuristic Algorithms

Today, balancing energy supply and demand due to limited resources has become one of the most important concerns of governments. Therefore, researchers are trying to be able to manage energy consumption in different ways. One of the best methods to predict the future is data analysis. Among data analysis methods, time series analysis is one of the most widely used methods for predicting the future. Hence, this study presents a recognized time series pattern and value estimation by using classical multilayer artificial neural networks. Moreover, the proposed pattern and values will optimize through the intelligent optimization algorithms. This chapter collects 40 years of natural gas consumption data from US industries and then preprocesses and prepares statistical data. In the next step by using particle swarm optimization algorithms and colonial competition, the model will train individually. Then, the result will be obtained via analyzing the answers and investigating the success of the trained network in adapting and recognizing the time-series pattern. This result shows that the combination of the classical model and optimization algorithms is successful and significantly increases the accuracy of prediction. In addition, it indicates less error than the classical model.