Multiobjective optimization of a plastic injection molding process

It is well known that the PID regulator has been very successful and widely accepted for controlling industrial systems where one objective function is the performance criterion. The question is whether the same success can be achieved for more than one objective function using the basic PID controller, while preserving its simplicity for use by an operator. Here the authors explore the use of multiobjective optimization for the purpose of extending PID controllers to applications that require the optimization of multiple objectives. An algorithm is implemented that runs in real-time and provides immediate feedback to a process operator about the current operating point. It was applied to the tuning of a PID controller to meet multiple objectives for the plastication phase of an injection molding process. The algorithm was implemented on an injection molding machine and allowed to tune the controller on-line. Simulation and experimental results are presented and compared