A Formulation of Artificial Kansei Systems Based on Multi-agent Spaces Generated by Variable Neighborhood Models

Abstract

In this paper, we formulate a method of designing artificial Kansei system based on a variable neighborhood model. We use a multi-agent space in order to realize a non-verbal system which can respond to individual situations. We carried out some simulation experiments of the system in which the system generates and produces some color arrangement like web pages as Kansei products. We also conducted a questionnaire survey in order to confirm that the system generates Kansei products which properly reflect users' Kansei.