Computational and Applied Aspects of Multiperiod Stochastic Programming

多周期随机规划的计算和应用方面

• 批准号: 9215921
• 负责人: John Birge
• 金额: $ 19.5万
• 依托单位: Regents of the University of Michigan - Ann Arbor
• 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: 1992-08-01 至 1995-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9215921&HistoricalAwards=false
• 关键词: Computational; Applied; Aspects; Multiperiod; Stochastic

This project will develop methods for approximations and solution procedures for multiperiod stochastic programs. The project will continue to develop and use realistic models from a variety of areas, including manufacturing, finance, vehicle routing, and energy policy. The major goals are to obtain practical solutions with known error characteristics efficiently. These goals will be accomplished through the development of new decomposition and interior point methods for distributed processors, new bounds on the value of the stochastic solution, new techniques for bounding stochastic nonlinear programming problems, and new approximations using parametric solutions and continuous distribution approximations. These methods will be assessed with the ongoing development of the stochastic programming model library. Comparisons will consider alternative approaches with computational effort, error analyses, and values of information and model complexity.

该项目将开发多周期随机程序的近似方法和求解程序。 该项目将继续开发和使用来自各个领域的现实模型，包括制造、金融、车辆路线和能源政策。 主要目标是有效地获得具有已知错误特征的实用解决方案。 这些目标将通过开发分布式处理器的新分解和内点方法、随机解值的新界限、限制随机非线性规划问题的新技术以及使用参数解和连续分布近似的新近似来实现。 这些方法将通过随机规划模型库的持续开发进行评估。 比较将考虑计算量、误差分析以及信息值和模型复杂性的替代方法。