Numerical Solution of Large-Scale Stochastic Programming Problems

大规模随机规划问题的数值求解

• 批准号: 9005159
• 负责人: Andras Prekopa
• 金额: $ 3.66万
• 依托单位: Rutgers University New Brunswick
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1990
• 资助国家: 美国
• 起止时间: 1990-08-01 至 1992-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9005159&HistoricalAwards=false
• 关键词: Numerical; Solution; Large; Scale; Stochastic

The study of large-scale systems and methods of computing their optimizations is an important area with many applications. Stochastic Programming is an important area in such studies. This activity will focus on development of stochastic programming problems. Large scale here means the number of constraints as well as the number of decision and random variables in the problem can be large. Both probabilistic (reliability type) constraints and penalties for deviations are incorporated into the models. The random variables are discrete in some of the models and continuous in others but stochastic dependence is allowed in all cases. In the case of continuously distributed random variables a multivariate numerical integral approximation technique, will be used. Possible applications include optimization of earthquake resistant structures, water level regulation in lake systems, planning in interconnected power systems, solutions of economic, finance, production problems.

大规模系统及其优化计算方法的研究是一个具有许多应用的重要领域。随机规划是这类研究的一个重要领域。这个活动将集中于随机规划问题的发展。这里的大规模意味着问题中约束的数量以及决策变量和随机变量的数量可能很大。概率（可靠性类型）约束和偏差惩罚都被合并到模型中。随机变量在一些模型中是离散的，在另一些模型中是连续的，但在所有情况下都允许随机依赖。在连续分布随机变量的情况下，将使用多元数值积分逼近技术。可能的应用包括抗震结构的优化、湖泊系统的水位调节、互联电力系统的规划、经济、金融和生产问题的解决方案。