Employing fuzzy logic and problem specific mutation methods to boost the performance of spectrum optimization via genetic algorithms

Abstract

This paper presents an improved method for determining the optimal filter (with respect to efficiency) to move a lamp from its "natural" position in color space to an arbitrary position in color space. Compared to a fixed parameter GA application employing the "chromosome smoothing operator" the use of fuzzy control of some GA parameters and application specific mutation methods leads to a substantial reduction in the number of function evaluations required, while maintaining the same overall level of solution quality.