A Method in Social Reasoning Mechanism for Intelligent Agents Using Fuzzy Inference

Abstract

Artificial agents should be socially intelligent in order to integrate with human society. Socially intelligent agents should be capable of reasoning about dependence relation, and we reconsider past research of dependence reasoning mechanism and integrate fuzzy inference with it to improve agents’reasoning ability. The simulation system consisting of agents within the social system is developed by incorporating the concept of trust and character of agents. These agents form a multiagent system made of three agents organized hierarchically. Fuzzy rules were implemented in our simulation system, and different experiments were performed.