The central objective of this work is the application of a method of a machine learning algorithm called DGM (Deep Galerkin Method), in order to solve an optimal abatement problem in carbon emissions formulated by Hambel, Kraft and Schwartz. To this end, the main results of stochastic control, the original formulation of the DGM and the work Optimal Carbon Abatement in a Stochastic Equilibrium Model in which the optimal abatement problem was formulated as a control problem were presented