A filter SQP algorithm without a feasibility restoration phase (Q843266)

The authors study the constrained optimization problem of a twice differentiable function. They begin by introducing sequential quadratic programming (SQR)algorithms and their use in solving this problem, as well as a short overview of the literature. The second section presents in detail the proposed filter sequential quadratic programming algorithm, including the derivation of the algorithm, properties of the iterate and the steps of the algorithm which include three loops. The third section studies the global convergence properties of the algorithm, where several theorems are presented and proven. The fourth section contains a series of numerical results carried out on previously studied test problems. The paper concludes with a list of useful references.