结合定量信息的定性仿真方法  

Method of Qualitative Reasoning with Quantitative Information

作　　者：吴瑞明[1,2] 刘豹[1,2] 关伟[3] 

机构地区：[1]中国矿业大学工商管理学院 [2]天津大学系统工程研究所 [3]北方交通大学智能交通研究所

出　　处：《中国矿业大学学报》1999年第4期361-364,共4页Journal of China University of Mining & Technology

基　　金：博士点基金

摘　　要：在定性模拟算法（ＱＳＩＭ）中，时间被表示为由一些抽象的特异时点组成的集合，这些特异时点本身不含有任何具体的时域信息．由于缺乏时域信息，人们常常无法对带有时间延迟特性的系统进行仿真，因此必须充分利用一切可以利用的信息．根据区间约束传播思想，在ＱＳＩＭ算法基础上，本文提出了一种充分考虑时域信息的半定量方法（Ｓ－Ｑ），它是对ＱＳＩＭ算法的一个改进．本文还通过对离散经济系统ＣＯＢＷＥＢ模型的定性模拟。In the QSIM algorithm, time is often expressed as the set composed of abstract time points of nonstandard , and the time points have no temporal information. Because of temporal information lacking, the systems with character of lag time cant be simulated. By taking the qualitative approach, all the information that can be used is useful. Based on the QSIM algorithm, a semiquantitative simulation method (short in SQ) is given according to the thought of interval constraint propagation. The SQ method is an improvement on the QSIM algorithm. At last, the SQ method is applied to simulate the COBWEB model, famous economic system with timelag. The simulation process based on nonlinear envelope function is discussed.

关 键 词：定性推理 定性仿真 结合定量信息 人工智能 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]