Explore chapters and articles related to this topic
ADMM for multiaffine constrained optimization
Published in Optimization Methods and Software, 2020
Wenbo Gao, Donald Goldfarb, Frank E. Curtis
Given a graph and edge weights , the (weighted) maximum cut problem is to find a subset so that is maximized. This problem is well known to be NP-hard [30]. An approximation algorithm using semidefinite programming can be shown to achieve an approximation ratio of roughly 0.878 [22]. Applying the Burer–Monteiro approach [7] to the max-cut semidefinite program [22] with a rank-one constraint, and introducing a slack variable (see 4.1.2), we obtain the problem
It is easy to verify that all subproblems have very simple closed-form solutions.