A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering in Design and Manufacturing Engineering, School of Engineering and Technology

Thesis (Ph.D.) - Asian Institute of Technology, 2007

The dissertation proposes a new aesthetic-driven evolutionary algorithm (EA) for user-centered jewelry design as an art form generator. The EA is hybrid between evolutionary strategy, evolutionary programming, genetic algorithm, and genetic programming. Jewelry art forms are represented using Iterated Function System (IFS) fractals and encoded in form of chromosomes. Multi-Gaussian mutation and modified arithmetic crossover are developed to increase diversity and productivity. In cooperation with the multiple parent system, it offers co-evolution, that good genes can be transferred from one generation to the next generation. In mapping IFS genotypes to fractal phenotypes, we propose to use IFS with uniform random probabilities in Random Iteration Algorithm for rendering fractals as source of variations and inspirations. Aesthetic algorithm is presented for measuring aesthetics of art forms as fitness function. It is validated by famous aesthetic theories, surveys of the universal aesthetic preference for human beauty and popular symbols. The results of statistical analysis are embraced in this study to fine-tune the aesthetic algorithm. The aesthetic algorithm consists of two steps. The first step, morphological fitness function is developed to quantify compactness and connectivity of art forms and to discard unfit art forms, before going to the next step. The second step, aesthetic fitness function quantifies aesthetics of art forms in terms of fractal dimension, largest Lyapunov exponent, image complexity, golden ratio, mirror symmetry, rotational symmetry, and logarithmic spiral symmetry. The proposed design system basically creates two-dimensional (2D) designs. Condensation sets applied in the combination with IFS fractals allows user to integrate any 2D shapes into the existing design for generating the new designs. The system is practical for jewelry industry. The examples of jewelry designs and physical prototypes created by the system are included. The dissertation contributes to several fields such as "Artificial Intelligence in Design", "Design Study", and "Aesthetic Computing" especially for non-functional design.