General-Purpose Modeling Environments for Discrete Optimization

9414487福雷尔本研究集中于识别和研究在设计语言、界面和算法时必须解决的困难的概念性问题，以解决广泛类别的组合优化问题。这项研究的目的是确定将离散优化技术转移给广泛用户的策略，并设计支持离散优化应用程序开发的通用语言和环境。研究的基础是两个系统设计元素--代数建模语言和代数工具包，它们已用于统计和连续优化问题的软件开发。这项工作将共同考虑组合优化通用软件开发中的表示和算法问题。该研究将对工程教育和美国工业具有直接的实用价值。美国的工业每天都在致力于解决本质上是组合的问题。这项研究的结果将促进显式模型的易用性和强大的优化技术来解决这些问题。行业还将受益于通过使用优化模型而不是启发式算法获得的高质量解决方案。从工程教育的角度来看，通用建模软件的开发将在大规模优化教学中发挥关键作用。这将允许学生制定、解决和分析实际的大型和复杂的模型，而不是处理通常较小和人为设计的教科书类型的问题。

9414487 Fourer This research focuses on the identification and investigation of difficult conceptual issues that must be resolved in designing languages, interfaces, and algorithms for a broad classes of combinatorial optimization problems. The research aims to identify strategies for transferring discrete optimization technology to a broad range of users and to design general-purpose languages and environment that support application development in discrete optimization. The foundation of the research is provided by two system design elements, the algebraic modeling language and the algebraic toolkit, that have been used in the development of software in statistics and continuous optimization problems. This work will jointly consider the issue of representation and algorithms in developing the general-purpose software for combinatorial optimization. The research will have direct practical value to engineering education and the American industry. American industries engage daily in resolving problems that are combinatorial in nature. The outcome of this research will facilitate the ease of use of explicit models and powerful optimization technologies to resolve these problems. Industry will also benefit from the quality solution that can be obtained from the use of optimization models as against heuristic algorithms. From the engineering education view point, development of a general-purpose modeling software will play a key role in the teaching of large-scale optimization. This will allow student to formulate, solve, and analyze realistically large and complex models instead of dealing with textbook type problems that are generally small and contrived.