USE OF ARTIFICIAL INTELLIGENCE ALGORITHMS AND GOOGLE COLAB FOR AUTOMATING THE CALCULATION OF MATERIAL PIECES INTO SPREADS

Abstract

Purpose. To study the features of interaction between an apparel technologist without programming skills and Artificial Intelligence (ChatGPT) for the automatic generation of Python code. The research is aimed at solving combinatorial optimization tasks regarding the rational selection of material pieces into spreads to minimize inter-pattern waste and end remnants within the framework of the Industry 4.0 production digitalization concept.

Methodology. The study utilizes a comprehensive approach including systems analysis methods to describe cutting-room process constraints, prompt engineering methodology to form queries for AI, and combinatorial optimization methods (the Knapsack Problem) to calculate options for combining fabric rolls using Python in the Google Colab environment.

Results. The features of using Artificial Intelligence as a decision support tool for automating complex technological calculations in apparel manufacturing were investigated. A methodology for the rational selection of material pieces (rolls) into spreads using AI was developed. Based on the developed prompt, a workable code was generated using ChatGPT. Using this algorithm allowed for the necessary calculation within a few minutes. Mathematical iteration ensured minimal end material remnants (from 0.01 m to 0.12 m per roll), which fully complies with the established initial constraint of no more than 0.15 m.

Scientific novelty. The concept of using prompt engineering as a no-code programming tool for solving specific tasks in cutting-room production was theoretically justified and practically implemented. The possibility of effectively using modern AI text algorithms as decision support systems in the light industry was proven.

Practical significance. The proposed methodology ensures high accessibility of complex computational algorithms for small and medium-sized enterprises that lack the resources to implement expensive specialized modules of commercial CAD systems. The developed prompt template is a universal tool that can be easily adapted for calculating any type of textile material and various assortment groups (lingerie, sportswear, or children's wear) by changing the input parameters in the Google Colab cloud environment. Implementing the research results into the production process allows for a significant increase in the material utilization rate, reduces the cost of finished products, and minimizes the impact of the human factor at the stage of cutting planning. Furthermore, the methodology has high potential for use in the educational process for training future specialists in the fashion industry.