Abstract: General interactive genetic algorithms allow only one user to participate in evaluation for candidate solutions, so their results can only reflect one users preference. But we should try to meet as many users needs as possible when designing products. To solve this problem, a multi-user interactive genetic algorithm (MUIGA) applicable for group design was proposed. And the detailed design of the three key modules-population initialization module, single-population module and multi-population module was given. Finally, the proposed algorithm and the general single-user interactive genetic algorithm were applied to a fashion design system, respectively. The experimental results were compared to validate the feasibility of MUIGA in group design.

Key words: interactive genetic algorithm, multi-user, population initialization module, single-population module, multi-population module, group design, fashion design system