Stochastic Simulation Optimization: An Optimized Approach

随机模拟优化：一种优化方法

• 批准号: 1233376
• 负责人: Chun-Hung Chen
• 金额: $ 26万
• 依托单位: George Mason University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2012
• 资助国家: 美国
• 起止时间: 2012-09-15 至 2016-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1233376&HistoricalAwards=false
• 关键词: Stochastic; Simulation; Optimization; Optimized; Approach

The research objective of this award is to develop efficient simulation-based optimization methodologies to enable fast-time simulation-based decision making under uncertainty. Simulation and Optimization are two most popular tools in operations research. However, the combination of simulation and optimization is still facing huge efficiency concerns. A decision maker is forced to compromise on simulation accuracy, modeling accuracy, and the optimality of the selected solution. The research in this award intends to seek for a seamless integration of simulation evaluation and optimization search with focus on efficiency. A key contribution involves a trade-off between allocating computational resources for searching the solution space versus conducting additional simulation replications for better estimating the performance of current promising solutions. This research will develop a new Optimized Simulation Optimization (OSO) method which intends to maximize the overall efficiency of simulation optimization.If successful, the results of this research will provide a set of fast simulation-based optimization methodologies. With such methodologies, a decision maker will be able to model complex, stochastic systems, while obtaining the optimal design very efficiently. The benefit of developing an efficient simulation-based methodology is that it offers unprecedented flexibility to address a wide variety of problems in different application contexts. Example applications include revenue management systems, transportation systems, and manufacturing systems. An open-source simulation optimization software package based on the new methodologies will be developed and be available at a dedicated web site for dissemination purpose. The software will be easily accessible and beneficial to industry practitioners and academic researchers. It will also be usable in several different courses to educate students about simulation-based decision making.

该奖项的研究目标是开发有效的基于仿真的优化方法，以实现在不确定性下基于仿真的快速决策。模拟和优化是运筹学中两种最常用的工具。然而，模拟和优化的结合仍然面临着巨大的效率问题。决策者被迫在模拟精度、建模精度和所选解决方案的最优性上妥协。该奖项的研究旨在寻求模拟评估和优化搜索的无缝集成，重点是效率。一个关键的贡献涉及分配计算资源之间的权衡搜索的解决方案空间与进行额外的模拟复制，以更好地估计当前有前途的解决方案的性能。本研究将开发一种新的优化仿真优化方法（Optimized Simulation Optimization，OSO），以最大化仿真优化的整体效率，如果成功，本研究的成果将提供一套快速的基于仿真的优化方法。 有了这些方法，决策者将能够模拟复杂的随机系统，同时非常有效地获得最佳设计。开发基于仿真的高效方法的好处在于，它提供了前所未有的灵活性，可以在不同的应用环境中解决各种各样的问题。示例应用包括收入管理系统、运输系统和制造系统。将开发一个基于新方法的开放源码模拟优化软件包，并在一个专门网站上提供，以供传播。该软件将很容易获得，并有利于行业从业人员和学术研究人员。它也将在几个不同的课程中使用，以教育学生基于模拟的决策。