Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Consider the "Complex" Quadratic minimization problem \begin{align} \min_{\mathbb{x}\in \mathbb{C}^{N \times 1}}~\mathbf{{x}}^H\mathbf{Q}\mathbf{x}-2~\Re{(\mathbf{x}^H\mathbf{b})}+1 \end{align}

$\Re(.)$ denotes the real part. Here $\mathbf{Q}$ is a $N \times N$ positive definite matrix. $\mathbf{b}$ is a $N \times 1$ vector. I am familiar with the technique of putting this problem in the real domain (where it becomes in $2N \times 1$ dimension) and then using the lagrangian technique to solve the resulting problem. I was looking for some analytic technique which would solve it in the complex domain itself. Applying lagrangian technique for complex vectors is also fine.

share|cite|improve this question
up vote 3 down vote accepted

For problems of this kind it is often convenient to treat the $\mathbf{x}$ and $\mathbf{x}^H$ as independent variables (in the end they are linearly related to real and imaginary part of $x$). So we want to minimize $$E= \mathbf{x}^H Q \mathbf{x} - (\mathbf{x}^H \mathbf{b} + \mathbf{b}^H \mathbf{x}) +1.$$

In this case as $E$ is a real function (as it should otherwise minimization does not make too much sense), the equations $$ \nabla_{\mathbf{x}} E =0 $$ and $$ \nabla_{\mathbf{x}^H} E=0$$ are complex conjugates of each other. So it is enough to solve one of these. We choose $$ \nabla_{\mathbf{x}^H} E = Q \mathbf{x} - \mathbf{b} =0$$ with the solution $$\mathbf{x} = Q^{-1} \mathbf{b}.$$

share|cite|improve this answer

Let $Q^{1/2}$ be a Hermitian square root of $Q$. Then by completing square, we get \begin{align*} x^HQx-2~\Re{(x^Hb)}+1 &=x^HQx-x^Hb-b^Hx+1\\ &=\|Q^{1/2}x - Q^{-1/2}b\|^2 + (1 - b^HQ^{-1}b). \end{align*} Hence the minimum occurs at $x = Q^{-1}b$ and the minimum value is $1 - b^HQ^{-1}b$.

share|cite|improve this answer
(+1) elegant, but i forgot to include that I was familiar with this solution also. I was looking for something in the lines of "Fabian"'s answer. – dineshdileep Jan 23 '13 at 8:48
I don't know if you are from a signal processing background, but it is really common place to use this technique you showed in minimum mean square error (MMSE) estimation theory. I feel happy to see that answer yet again. – dineshdileep Jan 23 '13 at 8:52
@user1551 Would you please perform one more step to elaborate, how did you equate the following ? $x^HQx-x^Hb-b^Hx+1=\|Q^{1/2}x - Q^{-1/2}b\|^2 + (1 - b^HQ^{-1}b)$ – kaka Feb 21 '15 at 7:07
@kaka $\|u\|^2 = u^Hu$. – user1551 Feb 21 '15 at 7:43

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.