Genetic algorithm with a Neuro-Fuzzy fitness function for optimal fuzzy controller design

The design of fuzzy controller is the key of fuzzy control system, while the core of fuzzy controller design lies in fuzzy rules, whose performance determines the control effect of fuzzy system. General fuzzy rules are obtained from expert experience, in which much subjectivity exists. In this paper, a fuzzy controller is designed by taking an intelligent cantilever beam as the research object. And a method using the genetic algorithm to optimize fuzzy rules is proposed and the genetic coding as well as the fitness function is confirmed. Finally, the simulation model of intelligent cantilever beam is built by Matlab/Simulink, and the vibration control effects of fuzzy controller optimized by genetic algorithm are compared with those un-optimized. The simulation results indicate that the vibration amplitude of intelligent cantilever beam has a significant decrease and the vibration decay rate has a significant increase after the fuzzy rules optimized.

Download Full-text

Neuro-fuzzy fitness in a genetic algorithm for optimal fuzzy controller design

The 2013 International Joint Conference on Neural Networks (IJCNN) ◽

10.1109/ijcnn.2013.6706739 ◽

2013 ◽

Author(s):

Oscar Castillo ◽

Abraham Melendez ◽

Patricia Melin ◽

Leslie Astudillo

Keyword(s):

Genetic Algorithm ◽

Fuzzy Controller ◽

Controller Design ◽

Neuro Fuzzy

Download Full-text

Reinforcement Hybrid Evolutionary Learning for TSK-type Neuro-Fuzzy Controller Design

IFAC Proceedings Volumes ◽

10.3182/20080706-5-kr-1001.01101 ◽

2008 ◽

Vol 41 (2) ◽

pp. 6530-6535 ◽

Cited By ~ 1

Author(s):

Yung-Chi Hsu ◽

Sheng-Fuu Lin

Keyword(s):

Fuzzy Controller ◽

Controller Design ◽

Evolutionary Learning ◽

Neuro Fuzzy

Download Full-text

Solving Vibration Control Problems Using Reduced Order Models for Flexible Structures: A Genetic Algorithm Approach

Volume 1B: 16th Biennial Conference on Mechanical Vibration and Noise ◽

10.1115/detc97/vib-3817 ◽

1997 ◽

Author(s):

Sourav Kundu ◽

Kentaro Kamagata ◽

Shigeru Sugino ◽

Takeshi Minowa ◽

Kazuto Seto

Keyword(s):

Genetic Algorithm ◽

Vibration Control ◽

Controller Design ◽

Fitness Function ◽

Flexible Structures ◽

Design Solution ◽

Reduced Order Models ◽

Reduced Order ◽

Fitness Value ◽

Genetic Algorithm Approach

Abstract A Genetic Algorithm (GA) based approach for solution of optimal control design of flexible structures is presented in this paper. The method for modeling flexible structures with distributed parameters as reduced-order models with lumped parameters, which has been developed previously, is employed. Due to some restrictions on controller design it is necessary to make a reduced-order model of the structure. Once the model is established the design of flexible structures is considered as a feedback search procedure where a new solution is assigned some fitness value for the GA and the algorithm iterates till some satisfactory design solution is achieved. We propose a pole assignment method to determine the evaluation (fitness) function to be used by the GA to find optimal damping ratios in passive elements. This paper demonstrates the first results of a genetic algorithm approach to solution of the vibration control problem for practical control applications to flexible tower-like structures.

Download Full-text