研究目的
Exploring an identification problem for nonlinear models and proposing a novel fuzzy identification method based on the ant colony optimization algorithm to improve approximation results and robustness.
研究成果
The proposed fuzzy identification method combines a modified cluster validity criterion index and ACO and OLS algorithms to establish the T-S fuzzy model with minimal error variance between the established model and the original nonlinear model. Examples show that the proposed method provides better approximation results and robustness than existing methods.
研究不足
The initial premise structure and consequent parameters might affect the final identification result. An insufficient number of rules will lead to an inappropriate premise structure and unreasonable modeling results.
1：Experimental Design and Method Selection:
A modified cluster validity criterion with a fuzzy c-regression model is used to find appropriate rule numbers of the Takagi-Sugeno fuzzy model. The ant colony optimization algorithm is then adopted to obtain the sifted initial membership function and the consequent parameters.
2：Sample Selection and Data Sources:
Input-output data pairs are used to partition into hyper-plane-shaped clusters for fuzzy modeling.
3：List of Experimental Equipment and Materials:
Not explicitly mentioned.
4：Experimental Procedures and Operational Workflow:
The improved ACO algorithm is used to find better initial premises and consequent parameters, followed by the use of an improved FCRM algorithm and the OLS method to obtain the final premise structure and consequent parameters.
5：Data Analysis Methods:
The mean square error (MSE) and average percentage error (APE) are used as performance indices to compare the established model with the original nonlinear model.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
