A predictor-corrector smoothing method for symmetric conic linear programming

A predictor-corrector smoothing Newton method is presented for solving the symmetric conic linear programming (SCLP). The proposed algorithm reformulates the SCLP as a nonlinear system of equations and then applies Newton´s method to the perturbation of this system of equations. It is shown that the algorithm is globally and locally quadratically convergent under suitable assumptions. Numerical results indicate the effectiveness of the algorithm.