Solutions in ML and Optimization

AIControl specializes in the boundary between mainstream traditional inferential statistics, closed loop optimal control, ML and AI. The intent is to provide a synthesis of these methods to solve the user’s problem in a simple and efficient manner. Current work is underway to facilitate advances in the Refining and Petrochemical fields with the American Fuel and Petrochemical Manufacturers (AFPM).

Specific Projects

 
  • Review AI weaknesses and strengths as an aid to improving Safety and dealing with near miss incidents in Operations

  • Reinforcement Learning (RL) and NN assisted RL (DRL) are a practical paradigm in some applications. We explore the many ways that DRL & RL can supplement Model Predictive Control in production processes in real time.

  • Model Predictive Control (MPC) is an established and proven technology for multivariable constrained control and economic optimization of process units. Machine Learning can help to extend the range of inferentials and incorporate model nonlinearities into the linear MPC structure.