混合智能系统R-FC-DENN及其实现  

Hybrid Intelligent System R-FC-DENN and Its Implementation

作　　者：王刚[1] 黄丽华[1] 张成洪[1] 

机构地区：[1]复旦大学管理学院,上海200433

出　　处：《科技导报》2007年第11期69-73,共5页Science & Technology Review

基　　金：国家自然科学基金项目(70571016;70471011)

摘　　要：混合智能系统已成为人工智能的一个重要研究方向,本文依据认知心理学和模型集成理论,构建了集粗糙集理论、聚类理论、模糊逻辑理论、遗传算法理论、人工神经网络理论于一体的混合智能系统R-FC-DENN,该系统在数据的空间效能方面和算法的时间效率等方面比已有算法有较大的改进。以混合智能系统R-FC-DENN为例,基于Rosetta和Matlab 6.5平台,还开发了混合智能系统工具箱RFCDENN-Tool,为相关研究打下坚实的基础。The hybrid intelligent system has become an important research direction of artificial intelligence. Based on the cognitive psychology and aggregative model theory, a new hybrid intelligent system R-FC-DENN is proposed based on rough set, clustering theory, fuzzy logic, genetic algorithm and artificial neural network. By means of clustering and simplifying the original dataset, the system can enhance training efficiency and robustness. With the hybrid intelligent system R-FC-DENN as an example, a toolbox RFCDENN-Tool is developed based on Rosetta and Matlab 6.5, which may provide a basis for related researches.

关 键 词：混合智能系统 人工神经网络 粗糙集 模糊聚类 遗传算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]