知识和数据协同驱动的群体智能决策方法研究综述EI北大核心CSCD

群体智能(Collectire intelligence,CI)系统具有广泛的应用前景.当前的群体智能决策方法主要包括知识驱动、数据驱动两大类,但各自存在优缺点.本文指出,知识与数据协同驱动将为群体智能决策提供新解法.本文系统梳理了知识与数据协同驱动可能存在的不同方法路径,从知识与数据的架构级协同、算法级协同两个层面对典型方法进行了分类,同时将算法级协同方法进一步划分为算法的层次化协同和组件化协同,前者包含神经网络树、遗传模糊树、分层强化学习等层次化方法;后者进一步总结为知识增强的数据驱动、数据调优的知识驱动、知识与数据的互补结合等方法.最后,从理论发展与实际应用的需求出发,指出了知识与数据协同驱动的群体智能决策中未来几个重要的研究方向.

Knowledge-based and Data-driven Integrating Methodologies for Collective Intelligence Decision Making: A Survey

Collective intelligence (CI) shows promising application prospects. Current research methodologies of intelligent decision making for CI systems can be categorized as knowledge-based and data-driven methods, both showing inherent advantages and disadvantages. Therefore, we claim that integrating knowledge-based and data-driven paradigms offers a new and prospective research direction. In this paper, possible methods of this integration are systematically introduced, and all of these methods are classified into a framework level and an algorithm level. Specifically, the methods integrated in the algorithm level are further categorized as hierarchical and componentized methods. In the hierarchical taxonomy, neural network tree, genetic fuzzy tree, and hierarchical reinforcement learning are included. In the componentized taxonomy, knowledge enhanced data-driven, data optimized knowledge-driven, and complementary knowledge and data driven methods are introduced. Finally, several future research priorities on the knowledge-based and data-driven integrating paradigms are proposed for the considerations of theoretical development and application requirement.