Abstract: The simulations of human cognitive processes have always been an important research area in both artificial intelligence and computational cognitive science. Cognitive architecture splits human cognition into modules, and simulates human cognitive processes by interactions between these modules. ACT-R (adaptive control of thought-rational) is a representative theory in cognitive architecture area, as a hybrid cognitive architecture, it uses symbolic and sub-symbolic systems to simulate human cognitive processes. ACT-R theory has been widely used in some areas like cognitive tutors and cognitive agent. This paper summarizes the related research and development of ACT-R, and introduces system components of symbolic and sub-symbolic systems, module functions and applications of ACT-R theory.

Key words: computational cognitive science, cognitive architecture, adaptive control of thought-rational (ACT-R), cognitive tutors, cognitive agent

摘要： 模拟人类认知过程是人工智能与计算认知学的一个重要研究领域。认知体系结构将人类认知模块化，通过模块之间的相互作用模拟人类认知过程。理性思维的自适应控制系统（adaptive control of thought-rational，ACT-R）是认知体系结构研究领域中的代表性理论，作为混合型认知体系结构，ACT-R使用symbolic系统与sub-symbolic系统共同模拟人类认知过程。ACT-R的相关研究已经应用到智能导师、智能agent等领域，并得到了越来越多的关注。通过对ACT-R的发展与相关研究进行综述，分别介绍了symbolic与sub-symbolic的系统组成、模块功能，以及ACT-R的相关应用等内容。

关键词: 计算认知科学, 认知体系结构, 理性思维的自适应控制系统（ACT-R）, 智能导师, 智能agent