Abstract:

Objective In the context of the rapid development of group customization of garment, this paper aims to solve the current problem of regional size system establishment of group customization, optimize the group customization number sizing process, and improve the style unity of group orders.

Method The Monte Carlo method was adopted to simulate the regional group size measurement data, and the regional group size system (RGS) was established based on the K-means clustering analysis results of the simulation data and the size proportion coefficient of the enterprise initial size system (EIS) and the sizing process of the group order was realized by using RGS. The sizing results were evaluated by comprehensive fit and aggregate loss.

Results The scatter diagram of chest grith/garment length of 10 000 sets of data was simulated by Monte Carlo method. The scatter diagram distribution and frequency distribution statistics of simulation data and real data were consistent (Fig.4), suggesting that the simulated data from the simulation model could be used as the target data of the next experiment. The clustering center of the simulation data clustering analysis results was tested (Tab.2). Based on the clustering center and EIS, the regional group size system RGS was established through the size system establishment process. 19 Types of RGS were involved in the research (Tab.4). Each size contained 4 important primary dimensions, i.e. chest grith, garment length, across shoulder and sleeve length. It was found that 4 sizes determining the style of garment were mid waist, hipline, sleeve bicep, and cuff. The same circumference or width transverse parameter has 3 length index values to match, and there were more size combination. Compared with EIS, RGS established on the basis of simulation data demonstrated a significant effect on the overall coverage of regional population. The comprehensive fit results showed that RGS can cover most target groups (Fig.5). The size data of the orders containing 218 people of the enterprise were matched by EIS and RGS, respectively. The sizing process and the sizing results were suggested 1 person cut individually for special body type. When using EIS for matching sizes, the evaluation result of the fit degree was as follows: 108 people fedback with excellent fitting (48%), 102 people good, 1 person general, and 6 people not ideal. When RGS was adopted to match the size of the order, and the fit degree evaluation results showed that 211 people returned excellent feedback, 4 people good, and 2 people general, with the proportion of excellent of 97.2%. The excellent conversion rate of RGS is 47.4% (Tab.5).

Conclusion The Monte Carlo method was adopted to simulate the regional size data accumulated by enterprise orders, and the enterprise size database was established. The simulation data reached the expected simulation goal, and the regional group size system could be optimized by using the simulation data. RGS establishment is shown to increase the coverage of target population, increase the fit degree of clothing, effectively improve the uniformity and consistency of garment customization, and thus reduce the probability of garment repair. The sizing mechanism considers the multidimensional size proportion of human body, changes the method of relying on a single dimension for the sizing process, and carries on the size matching through the proportion of different dimensions. The sizing process can effectively distinguish and classify the body size, and match the individual size with the size system. The sizing mechanism can provide theoretical basis for enterprise digital sizing process.

Key words: garment group customization, data simulation, K-means clustering algorithm, Monte Carlo method, garment sizing process