Interactive Evolutionary Computation utilizing subjective evaluation and physiological information as evaluation value

摘要

Interactive Evolutionary Computation (IEC) is known as an efficient method to create media contents suited with a user. However, user's fatigue in consecutive evaluation task, especially for time-sequential media contents, remains as a problem. Many previous studies have tried to solve this problem from algorithmic improvements. From different point of view, IEC methods using physiological information have been proposed. By mixing conventional IEC and IEC using physiological information, this paper proposes IEC method that utilizes both of user's subjective evaluation and physiological information to obtain evaluation value for a solution candidate. As a concrete system, IEC utilizing user's subjective relaxation feeling and heartbeat interval for composing relaxation musical piece suited with the user was constructed. To investigate the efficacy of the proposed method, comparing experiment with conventional IEC as control was performed. Created representative musical pieces in both methods afforded same level of relaxation feeling. The proposed method was effective for reducing a load in subjective selection, however, it increased fatigue by restricting body movement.