QSIM算法的序列因果关系约束和目标搜寻策略研究

分析了定性推理QSIM算法在经济系统等复杂系统听应用的不足，提出了增加序列因果关系约束和基于目标搜索的策略，修改了QSIM算法，提高QSIM对复杂系统的建模和仿真能力。     

基于云模型融入定性信息的定量仿真方法

针对系统定量仿真中存在的定性信息无法充分合理使用、定量信息和定性信息不能融合的问题,提出了一种基于云模型的定性仿真建模方法,并将其融入定量仿真之中.首先给出了基于云模型的变量空间描述,然后提出了基于云模型的融入定性信息的定量仿真方法.给出了包括定性与定量信息的融合运算、定性信息的持续与周期变化等在内的难点问题的解决方法.最后通过一个复杂供应链系统仿真实例验证了该方法的可行性.结果表明,改进方法具有客观的表达不确定性信息的特点,能将定性信息有效地融合到系统的定量仿真之中.     

一种基于单元分解仿真算法一SRSIM

针对标准的QSIM算法对于复杂系统推理求解时会产生不可控分技问题,提出一个分解仿真方法-SRSIM算法.应用聚类分析方法将定性模型分解成单元,对单元分开单独应用QSIM推理,对单元间的关系通过约束传播使得子模型推理一致.这种分解仿真算法可大大降低QSIM算法的不可控分枝,提高系统的仿真效率.     

基于云模型以约束为中心的复杂系统定性建模

针对以约束为中心的复杂系统仿真中缺乏有效的不确定性信息描述方法,导致仿真中不确定性信息不能充分利用的问题,借鉴和采纳云模型的相关理论和方法,研究了以约束为中心基于云模型的复杂系统定性建模方法。首先提出了仿真中基于云模型的不确定信息表示方法、基于云模型和群体专家决策的量空间构建方法;然后给出了基于云模型的系统定性约束方程的构建方法;最后将提出的建模方法应用于一个敏捷供应链的建模实例。结果表明提出的建模方法具有客观表达不确定信息、将定性与定量信息在仿真中相互融合的优点。     

一种基于单元分解仿真算法——SRSIM

针对标准的QSIM算法对于复杂系统推理求解时会产生不可控分枝问题,提出一个分解仿真方法--SRSIM算法。应用聚类分析方法将定性模型分解成单元,对单元分开单独应用QSIM推理,对单元间的关系通过约束传播使得子模型推理一致。这种分解仿真算法可大大降低QSIM算法的不可控分枝,提高系统的仿真效率。     

基于定性定量知识的故障诊断

提出一种将检测中的定量信息定性化的方法,并应用QSIM的算法约束的概念。结合区间代数的计算规则,设计了定性与定量结合的诊断系统,缩小诊断空间,增大定性仿真在故障诊断中的应用范围。通过不断扩大先验故障模型,使系统具备一定的学习能力,并以压缩制冷系统为例进行了诊断,验证了系统的正确性。     

融合时空信息的短时交通流预测

为了准确描述交通流的时空演化过程并提高交通流短时预测的精度,融合时空交通流信息,即时间维度的交通流量信息和空间维度的路网耦合信息,构造基于GM（1,N）-Markov 链的组合预测模型。将预测路段与关联路段看作是一个灰色系统并对其进行灰关联分析,通过对灰关联度最低阈值的设定,实现了空间信息的深度挖掘和对无效信息的过滤清洗;利用多维GM（1,N）模型对预测点与强关联点作全局、系统的分析预测,并针对GM（1,N）对随机性较大的数列可能出现预测失效的问题,引入马尔科夫链对模型进行修正;利用VISSIM对模型进行仿真验证,分别以2 min、5 min、10 min为时间间隔进行仿真模拟,预测平均相对误差分别为9.30%、5.95%、3.20%,模型精度均为优,证实模型是有效的。     

基于分层着色Petri网的热轧带钢厂产线建模

本文介绍了分层着色Petri网（HCPN）基本原理,并结合局部事件表提出了为复杂制造系统建模的一种方法,此方法可以减少复杂系统建模的复杂性,也为仿真软件体系结构的模块化和层次化设计建立了良好的基础.文中还给出了HCPN在武钢热轧带钢厂产线仿真系统建模中的实现方法.     

支持Modelica模型的半物理仿真系统设计及关键技术

针对当前基于MATLAB/SIMULINK的半物理仿真（Hardware in Loop Simulation,HILS）系统中不支持多领域建模、联合仿真模型转换过程复杂且存在误差等问题,提出基于Modelica和建模仿真软件MWorks进行HILS的方法,分析其所涉及的建模方法、实时仿真接口和代码生成等关键技术,并以此为基础构建支持Modelica模型的HILS.以前起落架转向系统和起落架收放系统为应用实例,建立HILS平台,对该方法进行验证.仿真结果表明该方法可行,结果可信度高.     

基于IDEF1X和XML的网上城市信息建模方法

针对网上城市信息交换中的信息建模难度大的特点,提出了一种用IDEF1X和XML技术进行复杂系统信息建模的方法.     

Parallel qualitative simulation

Qualitative simulation is a rather new and challenging simulation paradigm. Its major strength is the prediction of all physically possible behaviors of a system given only weak and incomplete information about it. This strength is exploited more and more in applications like design, monitoring and fault diagnosis. However, the poor performance of current qualitative simulators complicates or even prevents their application in technical environments. This paper presents the development of a special-purpose computer architecture for the bestknown qualitative simulator QSIM. Two design methods are applied to improve the performance. Complex functions are parallelized and mapped onto a multiprocessor system. Less complex functions are accelerated by software to hardware migration; they are executed on specialized coprocessors.     

AN OPS5 IMPLEMENTATION OF QUALITATIVE REASONING ABOUT PHYSICAL SYSTEMS

This paper describes the implementation of WQMS, a general-purpose qualitative modeling environment that integrates the process-centered approach and the constraint-centered approach. A new simulation algorithm is presented in which, owing to this integration, some drawbacks of QSIM are overcome using a depth-first strategy in the assessment of the qualitative states and a reduced set of input variables. With this algorithm it is possible to achieve qualitative simulations in cases in which simulations by QSIM are not feasible, because in some complex systems it is possible neither to specify the initial conditions for all required variables nor to handle the amount of branching that arises with normal computer facilities. An implementation of modeling and simulation of two biological systems is shown; to the best of the authors' knowledge, WQMS is the only environment with which a qualitative simulation can be achieved for these systems.     

具有交互效应的多变量GM(1,$textbf{N}$)模型

针对一类具有交互效应的小样本系统建模问题,将相关因素序列的交叉项引入经典GM(1,N)模型的灰色作用量,构建交互效应GM(1,N)模型及其派生模型,以反映不同输入变量之间的交互效应对系统特征变量的影响,并通过实例验证交互效应GM(1,N)模型的有效性.结果表明:当相关因素序列的交互作用系数为零时,交互效应GM(1,N)模型退化为经典GM(1,N)模型;对于具有交互效应的系统建模问题,交互效应模型较经典模型具有更高的模拟和预测精度.     

Learning Qualitative Models of Dynamic Systems

The automated construction of dynamic system models is an important application area for ILP. We describe a method that learns qualitative models from time-varying physiological signals. The goal is to understand the complexity of the learning task when faced with numerical data, what signal processing techniques are required, and how this affects learning. The qualitative representation is based on Kuipers' QSIM. The learning algorithm for model construction is based on Coiera's GENMODEL. We show that QSIM models are efficiently PAC learnable from positive examples only, and that GENMODEL is an ILP algorithm for efficiently constructing a QSIM model. We describe both GENMOEL which performs RLGG on qualitative states to learn a QSIM model, and the front-end processing and segmenting stages that transform a signal into a set of qualitative states. Next we describe results of experiments on data from six cardiac bypass patients. Useful models were obtained, representing both normal and abnormal physiological states. Model variation across time and across different levels of temporal abstraction and fault tolerance is explored. The assumption made by many previous workers that the abstraction of examples from data can be separated from the learning task is not supported by this study. Firstly, the effects of noise in the numerical data manifest themselves in the qualitative examples. Secondly, the models learned are directly dependent on the initial qualitative abstraction chosen.     

Postdiction using reverse qualitative simulation

Postdiction is the task of finding the possible pasts of a physical system, given its model and current state. We present an appropriately modified version of Kuipers' QSIM algorithm to perform postdiction by reverse qualitative simulation. The necessary changes to the algorithm are explained and the closed world assumption is discussed in this new light. The new algorithm can be used for diagnostic purposes     

Modified Grey Model Predictor Design Using Optimal Fractional-order Accumulation Calculus

The major advantage of grey system theory is that both incomplete information and unclear problems can be processed precisely. Considering that the modeling of grey model (GM) depends on the preprocessing of the original data, the fractional-order accumulation calculus could be used to do preprocessing. In this paper, the residual sequence represented by Fourier series is used to ameliorate performance of the fractionalorder accumulation GM(1, 1) and improve the accuracy of predictor. The state space model of optimally modified GM(1, 1) predictor is given and genetic algorithm (GA) is used to find the smallest relative error during the modeling step. Furthermore, the fractional form of continuous GM(1, 1) is given to enlarge the content of prediction model. The simulation results illustrated that the fractional-order calculus could be used to depict the GM precisely with more degrees of freedom. Meanwhile, the ranges of the parameters and model application could be enlarged with better performance. The method of modified GM predictor using optimal fractional-order accumulation calculus is expected to be widely used in data processing, model theory, prediction control and related fields.      

A review of: “LOGIC AND PHILOSOPHY IN THE LVOV-WARSAW SCHOOL”, by Jan Woleński,Kluwer, Dordrecht (The Netherlands) and Boston, 1989. xiv + 364 pages.

Qualitative methodology plays an important role within computer simulation; modeling and analysis of complex systems require qualitative methods since humans think naturally in qualitative and linguistic terms. The critical interface for simulationists exploring qualitative simulation should rely on an unambiguous mathematical formalism or method with foundations in systems theory. Currently, many ad hoc formalisms exist for encoding uncertain or qualitative simulation knowledge; however, we have found that fuzzy set theory provides for a formalism where linguistic variables can be encoded as state, parameter, input and output information in the model. Fuzzy numbers, in particular, are useful when population statistics are unavailable—usually due to cost factors. We have constructed fuzzy simulation programs based on our C-based SimPack library and we use fuzzy simulation to hypothesize qualitative system models reflecting real system behavior, and to specify qualitative versions of systems.