Interactive, Semi-Automated, Lagrangean Relaxation and Decomposition Modeling System for Linear and Mixed Integer Programming

用于线性和混合整数规划的交互式半自动拉格朗日松弛和分解建模系统

• 批准号: 9160722
• 负责人: Robert Kimble
• 金额: $ 5万
• 依托单位: DANIEL H WAGNER ASSOCIATES INCORPORATED
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1992
• 资助国家: 美国
• 起止时间: 1992-01-01 至 1992-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9160722&HistoricalAwards=false
• 关键词: Interactive; Semi; Automated; Lagrangean; Relaxation

This project addresses the need of the applications specialist for advanced mathematical programming methods without the heavy investment of time which is normally required. It aims to provide advanced modeling aids for setting up Lagrangean Techniques which are usually indispensable when dealing with large complex problems. This will be accomplished by imbedding Lagrangean Modeling into an interactive environment under state-of-the-art systems such as GAMS and MPSX/370. The user will be guided in setting up models and associated Lagrangean Techniques without having to code the details. He will be led by the system and will be able to try out various options before committing to a final set of decisions (how to decompose, how to set the various parameters, how to select from a set of specialized solvers, how to guide the Lagrangean Dual solution procedures, etc.) Various algorithmic devices will be provided as needed, including: solving the Lagrangean Dual by subgradient optimization or dual ascent; providing an overall Branch and Bound shell; making available special purpose algorithms; modifying GAMS to take advantage of the decomposition; introducing advanced integer programming (logical processing, probing) and heuristics.

该项目解决了应用程序专家的需求 先进的数学规划方法，而无需 这是通常需要的时间。 它旨在提供 先进的建模辅助设置拉格朗日技术， 在处理大型复杂问题时通常是必不可少的。 这 将通过将拉格朗日建模嵌入到 在GAMS等最先进系统下的交互式环境 MPSX/370 用户将被引导建立模型， 相关的拉格朗日技术，而不必编码的细节。 他将由系统领导，并将能够尝试各种 在做出最终决定之前的选择（如何 分解，如何设置各种参数，如何从集合中选择 如何引导拉格朗日对偶解 程序等） 将提供各种算法装置， 需要，包括：通过次梯度求解拉格朗日对偶 优化或双重提升;提供总体分支和界限 shell;提供专用算法;修改GAMS， 利用分解;引入高级整数 编程（逻辑处理，探测）和计算机科学。