In the world, there are various things designed by designers such as websites, clothes, interior cars, etc. Designers will design things by making use of various knowledge such as design methods, tools and color schemes, but general users do not have such a knowledge.

In such a case, even if general users do not have expert design knowledge, they can make products with various technologies rerated on kansei information processing to support the design of various things. One of these technologies, there is an ""interactive evolutionary computation"" technique.

Then, let's briefly look at the mechanism of interactive evolutionary computation.


For example, when a user coordinates shirt, inner, bottoms, and socks, first, the computer will present several coordination proposals to the user. First, the computer will present several coordination proposals to the user.

Then, the user evaluates the proposed coordination plan based on that sensitivity and preference such as “I like this coordinate” or “That one is not so good”.

Finally, the evaluation information of the user is fed back to the computer, and the computer creates a new coordination proposal based on the user's evaluation and presents it to the user again. Repeat the above steps to generate coordinate that matches the user's sensitivity.

Just only choose his/her likes!

In the field of product development in recent years, product development that matches the user's sensitivity is important, and application of interactive evolutionary computation is expected. However, the interactive evolutionary computation has a problem that users' evaluation burden is high. We have been conducting research on improving the evaluation interface from the viewpoint of simplifying evaluation to reduce user's evaluation burden.

We have proposed the following two methods to date. The first method is a tournament-style evaluation in which the user answers which of the two designs presented by the computer is preferred. Another way is interactive taboo search which selects the favorite one from multiple design ideas.



While proposing these prototype systems, we have seen the widespread adoption of smartphones and the increasing use of tablet devices in recent years. Riding this trend, we have also developed an interactive evolutionary computation system for designing wristwatches based on an Android app.
For more details, click here. → Makeup，Ladies' Watch Design，Men's Watch Design．

Additionally, we are researching interactive evolutionary computation systems that not only capture the sensibility information of a single user but also gather the sensibility information of many users as votes, as well as systems that utilize users' gaze information for interactive evolutionary computation.
For more details, click here. → Voting-based Collaborative Design，Gaze Information-based System．

Related our works
【Invited lecture】Hiroshi Takenouchi， "Dynamic Computer Systems Using User Sensibility Information"， Information Processing, Vol.62, No.5, pp.d45-d69, 2021-04 (in Japanese).
【Academic Papers】Hiroshi Takenouchi, Masataka Tokumaru, Noriaki Muranaka, "Interactive Evolutionary Computation Using a Tabu Search Algorithm", IEICE Transactions on Information and System, Vol.E96-D, No.3, pp.673-680, 2013-03.