Mathematical Sciences: Exact Penalty Functions in Constrained Optimization

该项目旨在开发鲁棒信赖域方法来解决约束非线性优化问题。特别关注的问题包括鲁棒全局收敛理论，识别有限次迭代中最优主动约束的算法的表征，二阶校正方法和并行计算技术的数学理论。

This project is directed at the development of robust trust region methods for the solution of constrained nonlinear optimization problems. Problems of particular concern include a robust global convergence theory, the characterization of algorithms that identify the optimal active constraints in a finite number of iterations, second order correction methods and a mathematical theory for parallel techniques of computation.