Nash-based criteria for selection of Pareto Optimal PI controller

This paper presents an approach to PI controller tuning based on an appropriate tradeoff selection from the Pareto front. The paper provides a formulation based on multiobjective optimization that analyses the tradeoff among different performance indexes. The corresponding Pareto front are generated by means of the normalized normal constraint (NNC), a numerical optimization method. As all the points from the Pareto front are compromise solutions, there is a need to select a unique solution that provides a reasonable control system. The Nash bargain solution is used for this purpose. On the basis of this selection, controller tuning rules are provided for PI controllers and First-Order-Plus-Dead-Time (FOPDT) models with normalized dead-times from 0.1 to 2.0.