Semidefinite Relaxations and Total Least Squares in the Second Order Conic Setting

In this paper we extend the well know total least squares problem to the second order conic setting. The corresponding equivalent model is a fractional quadratic problem with a second order cone constraint. To solve this equivalent model an iterative algorithm is designed that has to solve a semidefinite optimization relaxation at each iteration.

Keywords:

Total least squares , Second order cone , Semidefinite Relaxation