AN INEXACT LEVENBERG-MARQUARDT METHOD FOR TENSOR EIGENVALUE COMPLEMENTARITY PROBLEM

摘要

In this paper, we consider the method for solving tensor eigenvalue complementarity problem due to its various applications in engineering and polynomial optimization. We formulate the tensor eigenvalue complementarity problem as semismooth equations involving an NCP function and propose an inexact Levenberg-Marquardt method to solve it. The inexact Levenberg-Marquardt method is shown to be both globally convergent and local Q-superlinear convergence. Finally, numerical experiments are reported to show the efficiency of proposed method.