Using Suggestion Information in Exchange Solutions between Users in Interactive Evolutionary Computation Creating Blended Juices

Making products fit the feelings and preferences of many users is one of the targets of Kansei engineering. At present, it is still challenging to achieve the target. Interactive Evolutionary Computation (IEC) is a candidate for the techniques of making the products. IEC employs evolutionary computations’ searching ability for the best and good solutions to a particular problem. Some previous studies have proposed IECs where multiple users participated in the search. In some of them, the user’s basic search is performed independently, and sometimes good solutions are exchanged between the users. This study aims to propose a method using users’ suggestions in the exchange as a good communication opportunity between users. This study tried to achieve the users’ suggestion in IEC in the creation of juice by blending several source juices. The tasting experiment through seven generations was conducted with fourteen participants. Significant increases in mean and maximum fitness values were observed from the initial to the final generations. Furthermore, qualitative analysis is performed based on the exchange states and the reason for the suggestion with a thought of positive computing.