Learning Adaptive Parameters with Restricted Genetic Optimization Method

Abstract. Mechanisms for adapting models, filters, regulators and so on to changing properties of a system are of fundamental importance in many modern identification, estimation and control algorithms. This paper presents a new method based on Genetic Algorithms to improve the results of other classic methods such as the extended least squares method or the Kalman method. This method simulates the gradient mechanism without using derivatives and for this reason, it is robust in presence of noise.

• Type: Conference
• Year: 2001
• Where: IWANN
• Authors: Santiago Garrido, Luis Moreno
• Keywords: Control Algorithms, IWANN 2001, Kalman Method, Neural Networks, Squares Method