Creation of Juices with Interactive Genetic Algorithm via Taste Communication between Users in Remote Locations

Creating media content suited to many users is crucial from the viewpoint of business and a situation with multiple users' participation. Interactive Evolutionary Computation (IEC) is known as a search method that finds good solutions; however, it is originally for each user. This study proposes an IEC that gathers evaluations of multiple users in remote locations and utilizes them in the search, and the search target is good blended juices. Each user's good solution is exchanged between the users as a taste-communication. We had the experiment to investigate the proposed IEC's effectiveness with ten participants in five pairs: two participants of the pair located in different locations. The experimental results showed the effectiveness with a significant increase in fitness values and a significant convergence in the search space. Furthermore, details of the exchange of the GA individuals were investigated by observing the pair's exchange and comparing the fitness values of the exchanged GA individuals.