Learning Automata based set-point weighted parameter for unstable systems

A simple method using Learning Automata is given for selection of the set point weighting parameter (b) in PI controllers for unstable first-order plus time delay systems.The parameter (b) is selected, based on the state transition probabilities to reduce the over shoot for a servo problem. The performance evaluation of the control action is carried out by using the reinforcement scheme based on reward-penalty and accordingly state transition probabilities are updated. The learning automata consider the updated state transition probabilities as input and tuning parameter (b ) as output. The responses of PI controllers without and with the set point weighting are compared. © 2014 IFAC.