A Stochastic NARX Neural Network to Investigate the Carbon Capture in the Plantations of Forests

Abstract

Fast-growing forests play a vital role in decreasing global warming and have an extensive capacity for carbon capture. Three variables involved in the model are the quantity of living biomass, the intrinsic growth of biomass, and a forestry fire that has burned the area. This study explored the impact of environmental and ambient humidity parameters on the dynamics of fast-growing forest plantations. The nonlinear autoregressive network with exogenous inputs (NARX) technique is used to study the dynamics of fast-growing forest plantations. For the assessment of our soft computing technique, we use the Runge-Kutta fourth-order approach as reference solutions. The results of our simulations are compared with the reference solutions. It has been concluded that our approach is superior to the state-of-the-art. Regression, fitness, and error histogram plots are graphically displayed for further illustration of the results.

Keywords

• Computational intelligence
• artificial neural networks
• machine learning
• Runge-Kutta order four