Title: Multiobjective optimisation of robust Takagi-Sugeno fuzzy neural controller with hybrid learning algorithm
Authors: Ammar Soukkou, A. Khellaf, Salah Leulmi
Addresses: Department of Electronics, University of Jijel, BP. 98, Ouled Aissa, Jijel 18000, Algeria. ' Department of Electronics, University of Ferhat Abbas, Setif 19000, Algeria. ' Department of Electrotechnics, University of Skikda, Skikda 21000, Algeria
Abstract: A novel tuning and rule reduction algorithm is introduced as a design tool for the development of knowledge-based fuzzy neural controllers. A Hybrid Approach to Fuzzy Supervised Learning Algorithm (HAFSLA) that combines the Multiobjective Genetic Algorithms (MGA), Gradient Descent (GD) method and Least Square Estimation (LSE) technique is used to obtain the optimised fuzzy models. Simulations demonstrate that the proposed robust control has successfully met the design specifications for efficiency and robustness, respectively.
Keywords: fuzzy neural networks; multiobjective genetic algorithms; MGA; hybrid learning; nonlinear PI/PD controller; Takagi-Sugeno fuzzy control; robust control; tuning; rule reduction; fuzzy modelling; gradient descent; least squares estimation; simulation.
International Journal of Modelling, Identification and Control, 2007 Vol.2 No.4, pp.332 - 346
Published online: 28 Dec 2007 *Full-text access for editors Access for subscribers Purchase this article Comment on this article