A dynamic system approach to quadratic programming problems with penalty method

Abstract

In this work, we propose a dynamical system (state space) model approach to find a unique minimum of quadratic programming (QP) problem,;, with equality constrained. The unique minimum of the optimization problem is also proved to be asymptotically stable equilibrium point of the state space model. To obtain the optimal solution of QP optimization problem, we seek the limit point of the solution of the state space model by using the transfer function rather than discretization scheme. The numerical results are shown that the applicability and efficiency of the approach by compared with sequential quadratic programming (SQP) method in three examples.