Stationary Stochastic Models
Stationary Stochastic Models
复制标题
平稳随机模型
DOI：
10.1057/jors.1991.174
复制
发表时间：
1991
影响因子：
3.6
通讯作者：
W. D. Ray
中科院分区：
文献类型：
--
作者：
W. D. Ray
系统动力学在OR社区中没有得到更多的普及的几个原因中，缺乏该领域的好书肯定是最重要的原因之一。埃里克·沃斯滕霍尔姆（Eric Wolstenholme）关于这一主题的新书论述了这一情况，我很高兴地说，他做得非常好。我完全可以想象，这一案文将成为那些希望对这一领域有所了解的人的标准切入点。系统动力学的基本思想是，感兴趣的情况可以建模为由水平（可测量的量）和这些量的变化率组成的连续系统。分析人员首先利用影响图开发图解模型，用于对系统进行定性推理，并由此开发用于定量分析的仿真模型。与该领域的许多狂热者一样，作者认为系统动力学是一门独立的学科；在书的引言中，他将系统动力学与运筹学和“软系统”区分开来。我怀疑大多数现代的OR专业人员会认为，系统动力学虽然有自己的技术和方法，但与各种其他方法一起很好地适应现有的OR方法。本书的前半部分由介绍基本理论和实践的几个章节组成，包括几个小规模的例子来展示如何建立和使用模型。如果限于此材料，这本书将提供一个很好的介绍主题，但一个人可能已经离开了如何在实践中使用系统动力学的一点理解。因此，以下各章的案例研究是对第一部分的极好补充。在从商业战略、采矿作业和国防分析领域选择三个主要案例研究时，作者强调了该方法的广泛适用性。我发现特别令人高兴的是，作者强调了定性和定量分析阶段的重要性，而不是在“硬”和“软”轴上采取特定的立场。除了最微不足道的情况外，在所有情况下，都需要使用软件包来构建仿真模型；目前有两个主要的竞争者，这是作者的信用，两者都以公平的方式进行了讨论。事实上，案例研究使用了不同的包（尽管没有对同一案例研究使用这两种包的直接比较）。如果对这本书有什么批评的话，那就是对使用系统动力学的方法方面进行更多的讨论是受欢迎的。特别是，如何设计模拟实验以及如何解释这些模拟结果的一些有问题和微妙的方面没有得到任何深入的处理。我还找到了倒数第二章，
Of the several reasons why system dynamics has not gained more popularity within the OR community, the lack of good books in the area must be one of the most significant. Eric Wolstenholme's new text on the subject addresses this situation and, I am pleased to say, makes an excellent job of it. I can well imagine that this text will become the standard entry point for those who wish to gain some understanding of this area. The basic idea underlying system dynamics is that situations of interest can be modelled as continuous systems composed of levels (measurable quantities) and rates of change of these quantities. The analyst first develops diagrammatic models employing influence diagrams, which are used for qualitative reasoning about the system and from which simulation models are developed to be used in quantitative analysis. Like many afficionados of the area, the author argues that system dynamics is a self-contained discipline; in the introduction to the book he distinguishes system dynamics from operational research, and'soft systems'. I suspect that most modern OR professionals will take the view that system dynamics, while having its own techniques and approaches, fits well into existing OR methodology alongside the variety of other approaches.The first half of the book consists of several chapters introducing the basic theory and practice, including several small-scale examples to show how models are built and used. If limited to this material, the book would have provided a good introduction to the subject, but one might have come away with the little understanding of how system dynamics is used in practice. The case studies in the following chapters are thus an excellent complement to this first section. In selecting the three major case studies from the fields of business strategy, mining operations and defence analysis, the author highlights the wide range of applicability of the approach. I found it particularly pleasing that rather than adopting a particular position on the'hard'versus' soft'axis, the author stresses the importance of both the qualitative and quantitative phases of analysis. In all but the most trivial cases, the use of a software package is required to build the simulation model; there are currently two leading contenders and it is to the author's credit that both are discussed in an even-handed way. In fact, different packages are used for the case studies (although there is no direct comparison of the use of both for the same case study). If there is a criticism of this book, it is that more discussion of the methodological aspects of using system dynamics would have been welcome. In particular, some of the problematic and subtle aspects of how to design simulation experiments and how the results of such simulations should be interpreted are not treated in any great depth. I also found the penultimate chapter,