Evaluation of Autistic Disorder using Artificial School Class Game

本文采用多智能体模拟的方法，从生态学的角度，而非个体的角度，模拟学生之间的心理互动，探讨校园欺凌的成因。这样的人工学校类模型主要是能够用来澄清从本地的代理人的相互作用出现的类的全球现象。此外，我们设计了人工学校类游戏，通过用人类玩家替换其中一个代理。为了提高智能体的行为特性，通过游戏对人类行为进行了实验测量。最后，将球员的行为与五因素人格模型的结果进行了比较。

In this paper, a multi-agent simulation mimicking psychological interactions of school students was executed to investigate the causes of school bullying from not personal but ecological viewpoints. Such an artificial school class model is chiefly able to be used to clarify global phenomena of the class emerged from local interactions of the agents. We designed, furthermore, the artificial school class game by means of replacing one of the agents with a human player. In order to upgrade agent’s behavioral characteristics, human behaviors were experimentally measured by playing the game. Eventually, the behavior of the players was compared to the results from the five factor model on personality.