Optimization Algorithms for Advanced Computer Architectures

研究的主要目标是在高性能计算机上开发可移植和可扩展的大规模非线性优化算法。该研究将以制造系统的优化、超导模拟以及气象学和海洋学参数估计等应用领域为指导。将研究四类优化问题：(1)大型无约束问题；(2)变量边界简单的大问题；(3)变量多、自由度小的非线性约束优化；(4)结构化非线性约束问题。该项目的长期目标是满足大规模优化对高质量算法和软件的需求。这项研究将包括对新算法的理论分析，以及与它们在先进计算机体系结构上的有效实现相关的计算问题的研究。

The main objective of the research is the development of portable and scalable algorithms for large scale nonlinear optimization on high performance computers. The research will be guided by areas of applications such as the optimization of manufacturing systems, the simulation of superconductivity, and parameter estimation in meteorology and oceanography. Four classes of optimization problems will be studied: (1) large unconstrained problems; (2) large problems with simple bounds on the variables; (3) nonlinearly constrained optimization with a large number of variables and few degrees of freedom; (4) structured nonlinearly constrained problems. The long-term goal of this project is to fill the demand for high-quality algorithms and software for large scale optimization. The research will encompass a theoretical analysis of new algorithms, as well as the study of computational issues related to their effective implementation on advanced computer architectures.