Institutes outside Greece
Interior-point methods for linear and mixed complementarity problems
Abstract
dc:descriptionAn infeasible interior-point method with backtracking line search is proposed to solve the linear complementarity and the general mixed complementarity problem. The standard nonlinear complementarity problem is a special case of the latter problem. The algorithm is based on a method for the general nonlinear programming problem. For this algorithm we prove global convergence from any strictly positive staring point, under minor assumptions. Numerical results are reported which demonstrate very good computational performance on large scale linear complementarity problems as well as on a standard set of mixed complementarity problems.In the case of the general nonlinear program, the Karush-Kuhn-Tucker conditions can be stated as a mixed complementarity problem. Computational results are also reported on a large collection of nonlinear programming problems.Finally, the infinite horizon alternating move duopoly method is studied. It is shown that the necessary and sufficient conditions for a symmetric Markov perfect equilibrium of the model can be formulated as a nonlinear complementarity problem. When only necessary conditions are considered, a linear complementarity problem results. The performance of the algorithm when started from different points, is investigated on a specific example of this model. This example is known to have at least two equilibria, one is an Edgeworth cycle and the other is of the kinked demand curve type. Some sensitivity analysis of the solution around these two equilibria is provided.
Degree
thesis:*- Grantor dc:publisher
- Institutes outside Greece
- Year dc:date
- 1996
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Simantiraki, Evangelia
- Σημαντηράκη, Ευαγγελία
Subjects
dc:subject × 10Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/52006