Pathfollowing and jumps in bilevel programming

From MaRDI portal





In this interesting paper a bilevel programming problem is considered. The lower level problem is a one-parametric optimization problem. The concepts of optimistic and pessimistic strategies are introduced according to the assumptions on the reactions of the follower in front of the actions of the leader. The paper is restricted to the optimistic approach. Assuming that the lower level problem belongs to the generic class of one-parametric optimization problems introduced by \textit{H. Th. Jongen, P. Jonker} and \textit{F. Twilt} [J. Optim. Theory Appl. 48, 141-161 (1986; Zbl 0556.90086)], the authors develop an algorithm that provides a way to compute an optimistic solution to the bilevel programming problem. This algorithm is based on pathfollowing and jumps methods, under the global validity of the Mangasarian-Fromovitz constraint qualification at local minimizers of the lower level problem. The authors recognize that the practicability of the algorithmic procedure, based on maximal connected components of the parametric set of local minimizers of the lower level problem, seems to be restricted to relatively small problems. However, they propose two suggestions to deal with the time-consuming difficulty and the storage needs. For the entire collection see [Zbl 0951.00064].











This page was built for publication: Pathfollowing and jumps in bilevel programming

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2712295)