Three-Dimensional Modeling Design Of The Vase Based On The Interactive Genetic Algorithms

Vase as a commonly used household adornment,its modelling diversity.With the great enrichment of material life,consumers not only consider beautiful appearance,but also consider personalized performance when buying vases.In order to adapt to the changing needs of users,customized production has gradually become a trend.Therefore,enterprises need to understand the personalized needs of users faster and respond to the design production quickly.In order to design a vase that customers are satisfied with more efficiently,the best solution is to let customers directly participate in the design of the vase.Interactive genetic algorithm can effectively assist customers in product design,through the way of genetic evolution skillfully integrate products and people’s ideas,and finally get high user satisfaction modeling.This paper proposes the improved interactive genetic algorithm was carried out on the vase modelling design,in order to meet customers’ personalized needs of the vase.The main work of this study is as follows:(1)Since the study of vase modeling based on interactive genetic algorithm needs a target object for interaction design,the bicubic Bezier surface modeling method is adopted to construct the vase model,and then Web GL and Three JS related technologies are used to realize the visualization and interaction of the vase model.Finally,the vase model was parameterized and the genetic code of the vase was determined,which included the characteristics of the bottle surface,height and texture,making it the research object of interaction design.(2)Interactive genetic algorithm needs human participation in the optimization process,in order to solve the problem of human fatigue and evaluation subjectivity.Therefore,this paper proposes a combination of user evaluation surrogate model and graphical interaction mechanism to improve interactive genetic algorithm.The evaluation surrogate model is constructed based on the historical user evaluation data.The construction process adopts KD tree and random forest algorithm respectively.The nearest neighbor algorithm of KD tree is used to search the fitness value of similar individuals quickly,and random forest uses multiple decision trees to predict the fitness value of individuals,so as to assist users to evaluate individuals,help users to quickly recognize individuals,and reduce the number of user evaluation.Since the population gradually converges in the late stage of evolution,it is difficult to achieve fine optimization.Therefore,a graphical interaction mechanism is introduced to allow users to graphically interact with historical evaluation individuals,increase the diversity of the population,and make the evolution direction of the population more in line with users’ expectations.(3)This paper develops a vase shape design system based on the improved interactive genetic algorithm,and compares the improved algorithm with the traditional interactive genetic algorithm.The experimental results show that the improved algorithm proposed in this paper not only effectively reduces user fatigue,but also improves the optimization ability of the algorithm.