Genetic algorithms (GA) for adaptable design

Abstract

This thesis discusses the process of implementing and testing of a creative web element design system using a combination of genetic algorithms and K-nearest neighbor classification. Classification will be used as a learning mechanism to give the system the ability to absorb quality measures in regards to visual aesthetics, and utilize this knowledge to evaluate generated designs. In addition, some basic concepts regarding web design will be covered, along with an introduction to artificial intelligence-based design. The results of the implementation will be discussed and compared to related state-of-the-art systems.