Path-Following Methods for Generalized Nash Equilibrium Problems

TL;DR

This paper develops a path-following method for generalized Nash equilibrium problems with state-dependent constraints, including PDE constraints, providing a new computational approach with practical applications.

Contribution

It introduces a novel path-following strategy for solving generalized Nash equilibrium problems with complex state constraints, including PDEs, extending previous theoretical results.

Findings

Proposed a penalization-based approximation scheme.

Established a path-following method for equilibrium problems.

Illustrated with PDE-based control examples.

Abstract

Building upon the results in [Hinterm\"uller et al., SIAM J. Optim, '15], generalized Nash equilibrium problems are considered, in which the feasible set of each player is influenced by the decisions of their competitors. This is realized via the existence of one (or more) state constraint(s) establishing a link between the players. Special emphasis is put on the situation of a state encoded in a possibly non-linear operator equation. First order optimality conditions under a constraint qualification are derived. Aiming at a practically meaningful method, an approximation scheme using a penalization technique leading to a sequence of (Nash) equilibrium problems without dependence of the constraint set on the other players' strategies is established. An associated path-following strategy related to a value function is then proposed. This happens at first on the most abstract level and is subsequently established to a narrower framework geared to the presence of partial differential equations in the constraint. Our findings are illustrated with examples having distributed and boundary controls - both involving semi-linear elliptic PDEs.