On the interactive evolutionary computation page,we explained that the computer designs what matches the user's sensitivity. However, in normal interactive evolutionary computation, only a single user's sensitivity can be used. Generally, in product development, it is important to create products customized to the sensitivity of a single user, but it is also important to analyze the sensitivity of hundreds of users and extract the knowledge that can be used for product development. We study a new system incorporating "multiple people voting design evaluation" for interactive evolutionary computation."

On the current Internet, systems that allow users to vote on many websites are spreading such as like button of Facebook and retweet of Twitter.

In this research, we propose a system that acquires the sensibility of many users by voting and uses it for design evaluation of interactive evolutionary computation.

We have proposed technique called Paired Comparison Voting (PCV) Method. This method adopts the interface that many users vote the favorite one from the two designs presented from the system via the Internet.


In this system, we aim to design a sign for a tourist spot through a voting format involving local residents. While it is important for the design to catch the eye of travelers, incorporating the opinions of the local community is also essential. In such cases, having a system that can optimize designs through voting could enhance the effectiveness of interactive evolutionary computation systems.

So far, a certain level of effectiveness for the pairwise comparison voting method has been demonstrated through basic experiments using prototype systems. In the past, we have also verified various effectiveness of a method called the "Multi-User Participatory Tournament Method," which conducts pairwise comparison evaluations in a tournament format, through numerical simulations and experiments with actual users. Currently, we are challenging ourselves with large-scale evaluation experiments involving around 100 participants in the voting process, as well as developing evaluation methods for real-world environments.

Related our works
[Academic Papers] Masayuki Sakai, Hiroshi Takenouchi, Masataka Tokumaru, "Design Generation Support System Using Digital Signage with Voting from Multiple Users", Journal of the Japan Society of Kansei Engineering, Vol.15, No.4, pp.503-511, 2016-08.
[International Conference] Hiroshi Takenouchi, Hiroyuki Inoue, Masataka Tokumaru, "Signboard Design System with Votes by Many People", 2014 IEEE Symposium Series on Computational Intelligence (SSCI 2014), ISIC2014 proceedings, pp.14-19, 2014-12 (Orlando, Florida, USA).