STRUCTURAL MODEL OF MULTICONTOUR OPTIMIZATION IN THE INTELLIGENT FOOTWEAR DESIGN SYSTEM
Authors:
1.
ZAO MOF "Parizhskaya kommuna"
Moskva, Russian Federation
Moskva, Russian Federation
2.
ZAO MOF "Parizhskaya kommuna"
Moskva, Russian Federation
Moskva, Russian Federation
3.
Rossiyskiy gosudarstvennyy universitet im. A.N. Kosygina (Tehnologii. Dizayn. Iskusstvo)
Moskva, Russian Federation
Moskva, Russian Federation
4.
KNITU
Type:
Article
Pages:
from 164
to 172
Status:
Published
Received:
14.02.2026
Accepted:
13.03.2026
Published:
05.05.2026
Language:
Russian
Keywords:
INTELLIGENT FOOTWEAR DESIGN, MULTICONTOUR OPTIMIZATION, GENERATIVE STRUCTURAL SYNTHESIS, PREDICTIVE PRODUCTION ANALYSIS, COMPUTER VISION, MACHINE LEARNING
Abstract:
This study presents and experimentally validates a structural model of multicontour optimization in an intelligent footwear design system. The proposed methodology integrates generative neural network models, materials optimization, and predictive production analysis into a unified architecture of engineering and technological production preparation. The design process is formalized as a multicriteria optimization problem with an integral performance function that accounts for design time, material consumption, defect rate, and the accuracy of geometric reproduction of the finished product. To evaluate the geometric accuracy of transferring the digital model into the physical product, a volumetric shape similarity coefficient was employed, calculated through comparison of the CAD model and the 3D scan of the finished footwear. Experimental validation based on a paired sample of models demonstrated statistically significant improvements compared to the conventional approach: a 50% reduction in design time, a 25% decrease in material waste during cutting, a 50% reduction in production defects, and an increase in shape accuracy from 86.4% to 94.1%. The obtained results confirm the effectiveness of integrating artificial intelligence technologies into the digital design framework and demonstrate the possibility of simultaneous optimization of temporal, material, and geometric characteristics without mutual performance degradation. The proposed methodology can be integrated into existing CAD systems and production lines of mass-segment enterprises.
This study presents and experimentally validates a structural model of multicontour optimization in an intelligent footwear design system. The proposed methodology integrates generative neural network models, materials optimization, and predictive production analysis into a unified architecture of engineering and technological production preparation. The design process is formalized as a multicriteria optimization problem with an integral performance function that accounts for design time, material consumption, defect rate, and the accuracy of geometric reproduction of the finished product. To evaluate the geometric accuracy of transferring the digital model into the physical product, a volumetric shape similarity coefficient was employed, calculated through comparison of the CAD model and the 3D scan of the finished footwear. Experimental validation based on a paired sample of models demonstrated statistically significant improvements compared to the conventional approach: a 50% reduction in design time, a 25% decrease in material waste during cutting, a 50% reduction in production defects, and an increase in shape accuracy from 86.4% to 94.1%. The obtained results confirm the effectiveness of integrating artificial intelligence technologies into the digital design framework and demonstrate the possibility of simultaneous optimization of temporal, material, and geometric characteristics without mutual performance degradation. The proposed methodology can be integrated into existing CAD systems and production lines of mass-segment enterprises.
Keywords:
INTELLIGENT FOOTWEAR DESIGN, MULTICONTOUR OPTIMIZATION, GENERATIVE STRUCTURAL SYNTHESIS, PREDICTIVE PRODUCTION ANALYSIS, COMPUTER VISION, MACHINE LEARNING
INTELLIGENT FOOTWEAR DESIGN, MULTICONTOUR OPTIMIZATION, GENERATIVE STRUCTURAL SYNTHESIS, PREDICTIVE PRODUCTION ANALYSIS, COMPUTER VISION, MACHINE LEARNING
