Research On Product Multi-objective Image Form Evolutionary Design

Emotional satisfaction of product are considered by consumers. Product that embody consumer’s potential emotional needs and preferences are designed through kansei engineering, psychology, design, intelligent information, etc. It’s becoming increasingly important.In this paper, the overall process of analyzing product image form evolutionary-design is starting point. It includes four stages, which are determinating multi-objective image, defining and parameterizing study samples, establishing product image form evaluation systems, creating product image form evolutionary design system. The overall process is the main line in this study. In the first stage, technical methods of determination multi-objective image were studied, including cluster analysis, principal component analysis and multidimensional scaling. Taking automobile as case study,"gorgeousness","strength","steady","affinity","loveliness" and "streamline"6target image were determined by SPSS software through K-Means cluster analysis. In the second stage, technical methods of parameterization samples were studied, including curve control, parameter model and spectral analysis. Expression, disintegration, parameterization of automobile styling were discussed. Critical control-point coordinate-values were located at main styling line through curve control method, and as study sample parameters. In the third stage, technical methods of establishing product image form evaluation were studied, including fuzzy clustering analysis, quantification-Ⅰ and artificial neural network. Artificial neural network includes BP neural network, radial basis neural networks and fuzzy neural network. Establishment of input layer, basic neurons and output layer of fuzzy neural network evaluation system were proposed. Automobile image form evaluation systems were established. In the fourth stage, technical methods of product single-image form evolutionary design were studied. Four type evolutionary algorithms were summarized, including genetic algorithm, swarm intelligence algorithm, interactive evolutionary algorithm and hybrid algorithm. Model of product multi-image form evolutionary design was presented. Technology that is based on NSGA-Ⅱ algorithm of product multi-image form evolutionary design was proposed. Automobile multi-image form evolutionary design interactive system was exploited by MATLAB. The results presented that the overall process in this paper show is feasible. Model and methods of product multi-image form evolutionary design are effective.