Color schemes or color combinations are widely used in printed materials and their digital versions, including web pages. These schemes convey information and create aesthetic appeal. Consequently, graphic designers have to strike a balance between accessibility and aesthetics with the color schemes they select. However, certain color schemes make the represented information confusing when viewers have color vision deficiencies (CVDs), which may lead to decrease in discriminability between used colors. What we will call the color-scheme-recoloring problem is the difficulty of optimizing a color scheme so as to preserve its original design and yet be simultaneously discriminable for all viewers, with or without CVDs. To address the problem, we formulated it within a framework of a fuzzy constraint satisfaction problem, a technique of artificial intelligence, and developed a tool for graphic designers. By utilizing the practical color co-ordinate system (PCCS), we introduce concepts used in the field of graphic design. This work combines pre-publication recoloring techniques for CVD viewers with elements of artificial intelligence, psychophysics, and psychology.