ADMM (Alternating direction method of multipliers) for an optimization problem with non-separable term (multiplication of two variables)?

68 Views Asked by At

I'm looking for an optimization algorithm that solves a problem with a multiplied term. For example,

$$ \min_{x \in \mathbb{R}, z \in \mathbb{R}^{3}} \lVert xz + az + b\rVert_{2}^{2} $$ for given $a \in \mathbb{R}$ and $b \in \mathbb{R}^{3}$ constant.

I found a related question and answer in which ADMM is recommended.

So I've been scheming ADMM but it seems like it only works for a problem with "separable term" like $\min_{x, z} f(x) + g(z) \text{ subject to ...}$. And I don't think that my problem is not in the case.

Does it make sense as reported in the answer? If so, is there any theory for such problems (e.g. convergence analysis)?

1

There are 1 best solutions below

0
On BEST ANSWER

Choose $z = b$ and choose $x$ so that $x + a = -1$. Then the objective function value is $0$, so the optimization problem has been solved.