Convergence behavior of the central path for homogeneous
and self-dual cones
This is a small note resulted from the discussion among
Nesterov, Todd and myself. The result is that the central
path leads to a strict complementary solution if such
a solution exists. This result may be interesting to
many love-to-prove-superlinear-convergence researchers.
This result holds for all homogeneous self-dual cones.