Abstract for: AI-Driven System Simulation

This work presents an end-to-end methodology to use an integrated workflow facilitated in an open-source environment to use artificial intelligence and machine learning for estimating the parameters and then use the same setting for system dynamics simulations for decision making. The methodology provides a high-level overview of model development, data and preprocessing, training artificial intelligence models using machine learning, simulation, testing, and model longevity. This work also introduces the InsightConverter Python package which converts Insight Maker files to an XMILE format. This package is implemented into the methodology to help expand on cross-platform compatibility. Overall, the process is aided by a working example of a dynamical system using data alongside a support vector machine to estimate parameters