A modified SQP method and its global convergence (Q5899444)

The authors consider the minimization of a nonlinear function with nonlinear constraints. A modified sequential quadratic programming (SQP) method is proposed based on the subproblem presented in the paper by \textit{G. I. Zhou} [J. Glob. Optim. 11, 193--205 (1997; Zbl 0889.90135)]. The proposed algorithm has no requirement on the initial point and only needs to solve a quadratic programming subproblem. It overcomes the Maratos effect by solving an additional system of linear equations. Moreover, it avoids using a penalty function. Under some reasonable conditions, its global convergence is proved.