Compact Ant Colony Optimization Algorithms and Its Applications in Fuzzy-Neural Networks

• Main Authors: Chun-Yao Chen, 陳俊堯
• Other Authors: Chin-Ming Hong
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/17565298318409669794

碩士 === 國立臺灣師範大學 === 工業教育學系 === 97 === In this thesis, a compact ant colony optimization algorithm (CACOA) of fuzzy-neural networks is proposed. Traditionally, ant colony optimization algorithms solve discrete combinatorial optimization problems, and always need complicated operation procedures. Therefore, a continuous ant colony optimization algorithm is proposed for function approximation, nonlinear system modeling, and nonlinear system control. For function approximation and nonlinear system modeling, the weighting factors of the fuzzy-neural networks can be tuned through off-line learning procedure. For a class of multiple-input multiple-output (MIMO) nonlinear systems, the control scheme incorporates backstepping technique with the fuzzy neural networks, and the adjusted parameters of the fuzzy neural networks are tuned on-line via the CACOA approach. For state-feedback and output-feedback control, based on the direct adaptive control approach, a B-spline fuzzy-neural controller using CACOA is proposed to control a class of nonlinear systems. For the purpose of on-line tuning these parameters and evaluating the stability of the closed-loop system, an energy fitness function is included in the CACOA approach. The stability of the closed-loop system is analyzed by means of Lyapunov functions. In addition, in order to guarantee the stability of the closed-loop nonlinear system, a supervisory controller is incorporated into the CACOA-based B-spline fuzzy neural controller. Finally, the simulation results demonstrate the feasibility and applicability of the proposed method.