Uncertainty Management in Simulation-Optimization of Complex Systems : Algorithms and Applications
Uncertainty Management in Simulation-Optimization of Complex Systems : Algorithms and Applications
复制标题
DOI：
10.1007/978-1-4899-7547-8
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
G. Dellino;C. Meloni
中科院分区：
文献类型：
--
作者：
G. Dellino;C. Meloni
在生产、运输和物流、能源管理、金融、工程和应用科学中出现了一些优化问题；在所有这些情况下，管理者进行决策过程，通常受到不确定性的影响，因此最终结果可能是嘈杂的。几乎在商业和管理、政府、科学和工程的任何领域都有应用，因此迫切需要在不确定环境中支持决策的方法。这本书的目的是说明策略和方法，以说明计算机模拟所描述的复杂系统中的不确定性。在优化这些系统的性能时，考虑或忽略不确定性可能导致完全不同的结果；因此，不确定性管理是仿真优化中的一个重要问题。由于其广泛的应用领域，仿真优化问题已经被不同的社区用不同的方法和稍微不同的角度来解决。文献表明，已经开发了替代方法，也取决于应用程序上下文，没有任何行之有效的方法明显优于其他方法。使用流行的网络浏览器搜索关键词“Simulation Optimization”，返回大约20万页，而更集中的谷歌Scholar提供大约2.4万页，主要包含科学和技术文章、会议出版物、研究报告和学术手稿。显然，仿真优化是一个激发研究人员和处理实际问题的仿真实践者日益增长的兴趣的领域。作为文献中大量仿真优化工作的结果之一，专用的优化例程最近已被纳入几个商业仿真软件包中。这种流行的一个重要原因是，许多现实世界的优化问题太复杂，无法通过分析数学公式直接解决，而模拟模型避免了主要的简化（例如，可以考虑随机问题）。因此，优化和模拟社区的共同目标是开发方法来指导和帮助分析人员在缺乏可处理的数学结构的情况下产生高质量的解决方案。7
Several optimization problems arise in production, transportation and logistics, energy management, finance, engineering, and applied sciences; in all these contexts, managers carry on a decision-making process, which is usually affected by uncertainties, so the final outcome can be noisy. There are applications in virtually any area of business and management, government, science and engineering, so methodologies to support decisions in uncertain environments are urgently needed. This book aims at illustrating strategies and methods to account for uncertainty in complex systems described by computer simulations. When optimizing the performances of these systems, accounting for or neglecting uncertainty may lead to completely different results; therefore, uncertainty management is a major issue in simulation-optimization. Because of its wide field of applications, simulationoptimization issues have been addressed by different communities with different methods, and from slightly different perspectives. The literature shows that alternative approaches have been developed, also depending on the application context, without any well-established method clearly outperforming the others. An internet search using a popular web browser with the keyword “Simulation Optimization” returns about two hundred thousand pages, while the more focalized Google Scholar gives about twenty four thousand pages mainly containing scientific and technical articles, conference publications, research reports, and academic manuscripts. Clearly, Simulation-Optimization is a field that stimulates growing interest among researchers and simulation practitioners dealing with real problems. As one result of this great deal of work on Simulation-Optimization in the literature, dedicated optimization routines have been recently incorporated into several commercial simulation software packages. One important reason for this popularity is that many real-world optimization problems are too complex to be addressed directly through analytical mathematical formulations while simulation models avoid major simplification (eg, stochastic issues can be taken into account). Consequently, a common goal in both the optimization and simulation communities is to develop methods to guide and help the analyst to produce high quality solutions, in the absence of tractable mathematical structures. vii