We present an interior point method for the nonlinear complementarity problem which converges, whenever the problem has solutions, for any paramonotone operator (i.e., monotone and such that (F(x) -F(y), x-y) = 0 implies F(x) = F(y)). The iterative step consists of easily computable closed formulae,
Using an interior point method for the master problem in a decomposition approach
โ Scribed by J. Gondzio; R. Sarkissian; J.-P. Vial
- Publisher
- Elsevier Science
- Year
- 1997
- Tongue
- English
- Weight
- 764 KB
- Volume
- 101
- Category
- Article
- ISSN
- 0377-2217
No coin nor oath required. For personal study only.
โฆ Synopsis
We addres some of the issues that arise when an interior point method is used to handle the master problem in a decomposition approach. The main points concern the efficient exploitation of the special structure of the master problem to reduce the cost of a single interior point iteration. The particular structure is the presence of GUB constraints and the natural partitioning of the constraint matrix into blocks built of cuts generated by different subproblems.
The method can be used in a fairly general case, i.e., in any decomposition approach whenever the master is solved by an interior point method in which the normal equations are used to compute orthogonal projections.
Computational results demonstrate its advantages for one particular decomposition approach: Analytic Center Cutting Plane Method (ACCPM) is applied to solve large scale nonlinear multicommodity network flow problems (up to 5000 arcs and 10000 commodities). (~) 1997 Elsevier Science B.V.
๐ SIMILAR VOLUMES
provide an asymptotic analysis of a primal-dual algorithm for linear programming that uses modified search directions in the final iterations. The algorithm determines the search directions by solving the normal equations using the preconditioned conjugate gradient algorithm. Small dual slack variab
The control problem for an underactuated Lagrangian system is considered. A system of smooth nonlinear functions of the generalized coordinates is introduced into the treatment and the number of functions is equal to the number of generalized control forces. The aim of the control is to bring the sy