No label defined (Q4352285)

The authors present a method of solving the constrained extremum problem through converting it into separating of the sets in the image space. The problem is formulated as the generalized system \(F(x,y)\in \mathcal H\subset \mathbb R^\nu,\;x\in K \subseteq H,\;y\in Y\) with a cone \(\mathcal H,\;H\) a real Hilbert space, \(Y\) a parameter set. The impossibility of a generalized system means the condition \(\mathcal H\cap \mathcal K_y=\emptyset\), where \(\mathcal K_y:=F(K,y)\) is the image of \(K\) at \(y.\) The main problem reduced to impossibility is: \(\min f(x)\) subject to \(x\in R:=\{x\in K:g(x)\in \mathcal C\}\) with a convex and closed cone \(\mathcal C\subset \mathbb R^m.\) Applications to variational and quasivariational inequalities are shown. Results concerned with optimality conditions, saddle points, Lagrange multipliers, penalty method are presented, too. For the entire collection see [Zbl 0852.00018].