A second order continuous extragradient method with variable metric for solving equilibrium programming problems (Q1878772)

The author considers the problem of finding a point \(v^*\) such that \[ v^*\in\Omega, \quad \Phi(v^*,v^*)\leq \Phi(v^*,w), \quad \forall w\in\Omega, \] where \(\Omega\) is a closed convex set in \(\mathbb{R}^n\). The author suggests the following method for its solution \[ \begin{cases} \mu(t)G^{-1}(v(t))\ddot{v}(t)+\beta(t)\dot{v}(t)+v(t)= \\ \, \phantom{\mu(t)G^{-1}(v(t))\ddot{v}(t)}= \pi_{\Omega}^{G(v(t))}[v(t)-\gamma(t)G^{-1}(v(t))\nabla_w\Phi(u(t),u(t))],\\ \, u(t)= \pi_{\Omega}^{G(v(t))}[v(t)-\gamma(t)G^{-1}(v(t))\nabla_w\Phi(v(t),v(t))],\\ \, v(0)=v_0, \quad \dot{v}(0)=v_1, \quad t\geq0, \end{cases} \] where \(v_0\), \(v_1\) are any fixed points of \(\mathbb{R}^n\), \(\mu(t)>0\), \(\beta(t)\), \(\gamma(t)\geq0\) are parameters of the method, and \(G(v)\) is a given symmetric positive-definite matrix depending continuously on \(v\). The author proves the convergence of the method.