Automotive Styling Design Optimisation Based on Elliptic Fourier and Response-NSGA-II

Abstract

In order to solve the problem of how to establish the mapping relationship between the optimisation objectives and design parameters in automotive styling design optimisation and the optimal solution selection of Pareto solution set, an optimisation model combining elliptic Fourier descriptor, response surface analysis and NSGA-II algorithm is proposed. Firstly, the model uses the TF-IDF algorithm to extract consumers' affective preferences for car styling design, employs elliptic Fourier descriptors to quantify car styling, and reduces dimensionality through principal component analysis techniques. Secondly, response surface analysis is used to establish the fitness function between the optimisation objective and principal components, which is applied to the NSGA-II algorithm for multi-objective optimisation to generate the Pareto solution set. Finally, response surface analysis is used again to obtain the inter-correlation between the optimisation objectives to determine the optimal emotional preference combinations, and the unique optimal solution is filtered from the Pareto solution set based on the emotional preference combinations. The method is validated by its application in the optimisation of automotive styling design, successfully overcoming the plague of multiple covariance of the least-squares method in non-linear problems, and providing a clear criterion for the NSGA-II algorithm to select the optimal solution, which ensures the uniqueness and superiority of the optimisation results.