A smoothing Newton method for a type of inverse semi-definite quadratic programming problem (Q953410): Difference between revisions

The authors consider a cone-constrained optimization problem with a quadratic objective function. Instead of solving the problem directly, the authors present the smoothing Newton method to solve its dual problem. The global convergence is proved and the quadratic convergence rate is established. The effectiveness of the method is supported by some numerical results carried out in Matlab.