A conical algorithm for solving a class of complementarity problems (Q1089269): Difference between revisions

The concave complementarity problem: \(x\geq 0\), \(y=w(x)\geq 0\), \(x^ Ty=0\) \((x\in R^ n\), \(y\in R^ n\), \(w: R^ n\to R^ n\) a concave mapping) is converted into a concave minimization problem under convex constraints and solved by a conical algorithm for concave minimization. A special feature of this concave minimization problem is that its optimal value is equal to zero. This feature is exploited to devise simpler rules for bounding and fathoming than in the general concave minimization problem. The method solves the problem whenever it is solvable.