Running Shoes Design Support System Using Interactive Evolutionary Computation

Abstract

We propose running shoes design system using various interactive evolutionary computation interfaces. We proposed interactive genetic algorithm with tournament evaluation and interactive tabu search. And we inspected the effectiveness of these methods by simulation using evaluation function with imitation of human's Kansei. As a result, the evolution performance of the interactive genetic algorithm with the tournament evaluation was equal to interactive genetic algorithm with conventional 5 or 10-stage evaluations. Furthermore, the evolution performance of interactive tabu search was superior to interactive genetic algorithm. However, the effectiveness by the evaluation experiment for the real user is not shown. Therefore we built the system which designed running shoes by putting each part of running shoes together.