Wenhao Fu, lecturer, Suzhou University of Science and Technology

Yafeng Liu, Associate Professor

SQP-type method for constrained optimization with least constraint violation

15:30-16:30 October 17 (Tuesday), Z311

In this talk, we mainly introduce the Sequential Quadratic Programming (SQP) method for constrained optimization with the Least Constraint Violation (LCV). We call an LCV solution if it is a minimizer of the objective function among the set of stationary points for a certain measurement of the constraint violation, and present  some SQP-type algorithms to identify it. Under some boundedness assumptions, a globally convergent algorithm for Fritz-John points in nonlinear semidefinite programming is presented. Moreover, by utilizing the LCV solution of the quadratic programming subproblem, rapid convergence for SQP method in the field of nonlinear programming is analyzed. Finally, a penalty-free method for D-stationary points is shown. Numerical results for each algorithms are shown to demonstrate the effectiveness.

报告人简介：付文豪，讲师，2021年博士毕业于苏州大学，师从陈中文教授。主要研究领域为非线性规划和非线性半定规划的SQP型方法。获江苏省"双创博士"人才项目，发表SCI论文3篇。