GENERATIVE ADIVERSARIAL NETWORKS TO VALIDATE DATA IN DISCRETE EVENT SIMULATION

Computer validation is essential for the results of the simulation to be effective in the real system. The literature suggests the use of statistical techniques for this comparison, however, statistical assumptions may be violated. Thus, Generative Adverse Neural Networks (GANs) are an alternative since they adapt to any data. The work aims to use GANs to generate synthetic data from the real data and use the discriminator to discriminate real from simulated outputs. Five statistical distributions were trained and 10,000 distributions with the same characteristics were submitted to verify the power test. In addition, a real case of a Discrete Event Simulation in a hospital that assists patients with suspected COVID-19 was applied to the new validation technique. The results show that GANs are effective in discriminating data and can help in the validation of computer models.