Abstract:
Aiming at the shortcoming that the parameters of support vector machine (SVM) are not easily chosen, this paper utilizes the improved genetic algorithm (IGA) to optimize the parameters of SVM. By using generation gap selection and alterable crossover rate, the IGA algorithm can ensure that the present individuals of best adaptiveness are continuously passed to the next generation. Moreover, the optimized object is easily stabilized during the later stage of evolution, and the calculating efficiency can be raised. Finally, the proposed algorithm is applied to the fault diagnosis of acetic acid azeotropic distillation column. The simulation results show that this algorithm can obtain better classification result and faster training rate than the standard GA-SVM algorithm.