Genetic algorithm with knowledge-based encoding for interactive fashion design

  • Authors:
  • Hee-Su Kim;Sung-Bae Cho

  • Affiliations:
  • Department of Computer Science, Yonsei University, Seoul, Korea;Department of Computer Science, Yonsei University, Seoul, Korea

  • Venue:
  • PRICAI'00 Proceedings of the 6th Pacific Rim international conference on Artificial intelligence
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

Evolutionary computation gives a great potential in several realworld problems as a powerful tool for optimization and classification, but there still remain a lot of obstacles to be applied to artistic domains. To overcome the shortcoming a variety of techniques have been proposed, and among them interactive genetic algorithm (IGA) is extensively studied in these days. IGA exploits the interaction with human in the course of evolution by taking his evaluation as fitness. In this paper, we propose an effective knowledge-based encoding scheme for IGA in a real-world application. This method has been applied to a fashion design aid system, which can reflect user's preference or emotion that is usually difficult to be expressed explicitly. To show that the proposed encoding scheme produces more realistic and practical design, an experimental study as well as a theoretical investigation with schema theorem has been conducted.