Consider the following natural exponential function:
$f\left( {\bf{x}} \right) = {e^{{{\bf{a}}^T}{\bf{x}} + \frac{1}{2}{{\bf{x}}^T}{\bf{Cx}}}}$
where $\bf C$ is a positive definite matrix. Is there any tight lower bound for this function?
Consider the following natural exponential function:
$f\left( {\bf{x}} \right) = {e^{{{\bf{a}}^T}{\bf{x}} + \frac{1}{2}{{\bf{x}}^T}{\bf{Cx}}}}$
where $\bf C$ is a positive definite matrix. Is there any tight lower bound for this function?
Guide:
Notice that the exponential function is an increasing function, hence the lower bound is the smallest when $g(x)=a^Tx+\frac12x^TCx$ attains the smallest value.
You can differentiate $g(x)$ to solve for the minimal value.