1: \begin{abstract}

2: In this work we study a special minimax problem where there are linear constraints that couple both the minimization and maximization decision variables. 

3: The problem is a generalization of the traditional saddle point problem (which does not have the coupling constraint), and it finds applications in wireless communication, game theory, transportation, robust optimization,  just to name a few. We show that the considered problem is challenging, in the sense that it violates the classical max-min inequality, 

4: and that it is NP-hard even under very strong assumptions (e.g., the objective is strongly convex-strongly concave). 

5: We then develop a duality theory for it, and analyze conditions under which the duality gap becomes zero. 

6: Leveraging the resulting duality-based relaxations,

7: we propose  a family of efficient algorithms named multiplier gradient descent (MGD), and study their convergence properties.

8: Finally, we test the effectiveness of the proposed algorithms in an application on adversarial attacks on network flow problems.

9: 

10: 

11: \vspace{1cm}

12: 

13: \noindent {\bf Keywords:} min-max problem, coupled constraints, duality theory, computational complexity, max-min inequality, first-order algorithms.

14: \end{abstract}

15: