Large population size IGA with individuals' fitness not assigned by user

User fatigue problem in traditional interactive genetic algorithms restricts the population size. It is necessary to maintain large population size in order to apply these algorithms to optimize complicated problems. We present a large population size interactive genetic algorithm with an individual's fitness not assigned by the user in this paper. The algorithm divides a population into several clusters, and the maximum number of clusters is changeable with the evolution and the distribution of the population. A user only evaluates one representative individual in each cluster, and others' fitness are estimated based on these representative ones. In addition, to assign a representative individual's fitness, we record time when the user evaluates it satisfactory or unsatisfactory according to his/her sensibility, and its fitness is automatically calculated based on the time. Finally, we apply the proposed algorithm in a fashion evolutionary design system, and compare it with other two IGAs each of which has one aspect, including the population size and the evaluation method, the same as the proposed algorithm. The experimental results validate its efficiency.