I have the following optimization problem:

\begin{align} \min_{x,s,\lambda,\gamma} \quad J(x)\\ f(x)=0\\ h(x,\gamma,s)\leq0\\ \lambda^\top Gx \leq s\\ \lambda^\top e=1\\ \lambda\geq 0\\ \|G^\top \lambda\|\leq \gamma \end{align}

where $x \in \mathcal{X} \subset \mathbb{R}^n$, $s \in \mathbb{R}$, $\mathcal{X}$ is a convex set, $f(x)$ and $h(x)$ are affine functions, $J(x)$ is a quadratic function, $e$ is a vector consisting of all ones.

I know that the second constraint is a bilinear one. Is there any method to reformulate this problem into a mixed integer program? I know that constraints (3)-(5) could be reformulated into constraint on pointwise minimum, if there was no last constraint.


There are at least a couple of ways to approximate the bilinear constraint with linear constraints, possibly using some binary variables in the process. I can't think of any way to linearize it exactly.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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