Improved Complexity Analysis of Full Nesterov–Todd Step Feasible Interior-Point Method for Symmetric Optimization

In this paper, an improved complexity analysis of full Nesterov–Todd step feasible interior-point method for symmetric optimization is considered. Specifically, we establish a sharper quadratic convergence result using several new results from Euclidean Jordan algebras, which leads to a wider quadratic convergence neighbourhood of the central path for the iterates in the algorithm. Furthermore, we derive the currently best known iteration bound for full Nesterov–Todd step feasible interior-point method