Optimization of a problem with quadratic objective and quadratic constraints and alternative formulations

56 Views Asked by At

Suppose we have $A \in \mathbb{R}^{m \times n}$, we want to minimize

\begin{align} \min&\quad\|Ax\|_2^2 \\ \text{s.t.}&\quad x^\top C_1x = 1, \\ &\quad x^\top C_2 x = 1, \\ &\quad Dx > 0 \end{align}

Is there any optimization method to handle this?

Actually, the first and second constraints aims to force part of $x$ has the norm of $1$. The third constraint aims to force part of $x$ has the positive value.

Ignore the $Dx > 0$ constraint, can we formulate the optimization into a generalized eigenvalue problem or second-order cone programming?

Thanks in advance.