粒度模糊规则建模方法研究综述  

Review of the research of granular fuzzy rule-based modeling

作　　者：胡星辰 李妍[1] 陈紫健 李文涛[2] 申映华 刘忠[1] HU Xingchen;LI Yan;CHEN Zijian;LI Wentao;SHEN Yinghua;LIU Zhong(School of Systems Engineering,National University of Defense Technology,Changsha 410073,China;College of artificial intelligence,Southwest University,Chongqing 400715,China;School of Economics and Business Administration,Chongqing University,Chongqing 400030,China)

机构地区：[1]国防科技大学系统工程学院,湖南长沙410073 [2]西南大学人工智能学院,重庆400715 [3]重庆大学经济与工商管理学院,重庆400030

出　　处：《智能系统学报》2024年第1期22-35,共14页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金项目(62376279,61906204,72001032)。

摘　　要：本文旨在梳理粒度模糊规则模型中的主要研究及构建方法,并且进行系统分析与总结。粒计算是一种能模拟人类思维方式和求解复杂问题的新兴理论体系,以此为基础的粒度模型为复杂非线性系统的描述和问题求解探索了新的方向。粒度模糊规则模型将信息粒融入现有的模糊规则建模方法中,进行粒度级别的系统建模,以实现更高层次的数据分析与推理。本文简要介绍模糊聚类和模糊规则模型的基础知识;归纳了信息粒的构建方法,并讨论了相应的评估方法;总结了典型的粒度模糊规则模型的设计架构和优化方法。The purpose of this paper is to explore the main research and construction method for granular fuzzy rule model,and make a systematic analysis and summary of it.Granular computing is an emerging theoretical system that simulates patterns of human thinking and solves complex problems,and the granular model based on it explores a new direction for the description and problem solving of complex nonlinear system.Among which the granular fuzzy rule-based model incorporates information granule into existing fuzzy rule-based modeling methods to achieve granular-level system modeling for data analysis and inference at a higher level.This paper first briefly introduces the basics of fuzzy clustering and fuzzy rule-based models,next summarizes the construction method of information granules and discusses corresponding evaluation methods.Further,the design architecture and optimization method are summarized.

关 键 词：粒计算 信息粒 模糊规则 模糊C均值聚类 粒化和解粒化 粒度 原型 模糊集 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]