Algorithms, Applications and Theory of Mathematical Programming

数学规划的算法、应用和理论

• 批准号: 9101801
• 负责人: Olvi Mangasarian
• 金额: $ 23.87万
• 依托单位: University of Wisconsin-Madison
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1991
• 资助国家: 美国
• 起止时间: 1991-07-15 至 1995-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9101801&HistoricalAwards=false
• 关键词: Algorithms; Applications; Theory; Mathematical; Programming

Parallel algorithms will be developed for general nonlinear programs as well as for linear and quadratic programs. Emphasis will be placed on the recently proposed idea of constraint distribution among parallel processors as well as inexact and exact matrix splitting methods for the linear complementarity problem. Applications of mathematical programming to pattern recognition and neural networks will also be made, where the potential of optimization algorithms has not been fully realized. Particular attention will be given to medical diagnostic problems, some of which have been successfully solved by linear programming. Mathematical theory that is needed for the above algorithms and applications will be developed and sharpened. An important and promising area to be focused on is that of error bounds for various types of mathematical programs.

并行算法将被开发用于一般的非线性程序 以及线性和二次规划。 重点将 关于最近提出的约束分布的想法， 并行处理器以及不精确和精确矩阵分裂 线性互补问题的方法。 数学规划在模式识别中的应用 和神经网络也将被制造出来， 优化算法尚未完全实现。 特别 将注意医疗诊断问题，其中一些问题 已成功地解决了线性规划。 上述算法所需的数学理论， 将开发和改进应用程序。 一个重要而 有前途的领域是集中在误差界 用于各种类型的数学程序。