基于改进遗传算法的曲面拟合参数辨识

Spatial surface fitting is a complex highly nonlinear problem which is hard to solve. Existing surface fitting methods can barely figure out explicit curved surface expression. Viewing from the aspect of parameter identification, the parameters of coordinate translation, rotation and the standard curved surface equation are to be estimated. Genetic algorithm has particular advantages in this field. A series of modifications are made aiming at several defects of the classic simple genetic algorithm, and the improved algorithm is realized using MATLAB language. Both algorithms are used many times in double engineering examples, and the calculation results are compared. Applications and compares show that genetic algorithm can be well used in the field of spatial surface fitting, and the improved algorithm is better.