Quasi-convexity of the asymptotic channel MSE in regularized semi blind estimation

Abstract

In this paper, the quasi-convexity of a sum of quadratic fractions in the form ${\sum }_{i=1}^n\frac{1+c_i{x}^2}{{\left(1+d_{i}x\right)}^2}$ is demonstrated where $c_i$ and $d_i$ are strictly positive scalars, when defined on the positive real axis ${\mathbb{R}}^{+}$. It will be shown that this quasi-convexity guarantees it has a unique local (and hence global) minimum. Indeed, this problem arises when considering the optimization of the weighting coefficient in regularized semi-blind channel identification problem, and more generally, is of interest in other contexts where we combine two different estimation criteria. Note that V. Buchoux {\it et.al} have noticed by simulations that the considered function has no local minima except its unique global minimum but this is the first time this result, as well as the quasi-convexity of the function is proved theoretically.