Genetic Algorithm for Product Design Optimization: An Industrial Case Study of Halo Setting for Jewelry Design

Abstract

This paper proposes a new approach for the optimization of design parameters and the generation of design alternatives in product design process by using the generative design method (GDM). The proposed GDM is comprised of two stages: design optimization and design variation. This work focuses on developing a prototype system for jewelry design parameter optimization and shape variation. Using Genetic Algorithm (GA), the system will assist designers in creating halo ring designs with optimization of the number of halo gemstones and the gap size between them in the halo setting. Therefore, this system helps designers to avoid undesirable design iterations that cause waste of a great amount of designing time and effort. In GA the properties of the objects are described using chromosomes, where each chromosome is encoded with a set of jewelry design parameters. Uniform crossover and point mutation are utilized as genetic operation. Fitness function is derived from the mathematical relationships between the design parameters to calculate the gap size between halo stones. The empirical results indicate that the proposed approach is applicable for product design optimization. The proposed design system can reduce time for setting halo gemstones surrounding a center gemstone in comparison to manual setting using the trial-and-error method. Using this generative design technique, the proposed design system can automatically generate various design alternatives of halo rings with the set of optimized design parameters, such as center gemstone size, halo gemstone size, number of halo gemstones, and gap size. The experimental results of halo ring setting are also provided in this paper.