Solving convex (and linear) complementarity problems by projection methods (undamped Newton)
A recent approach for solving the Linear Complementarity Problem (LCP) has been the solution of an equivalent system of piecewise linear equations through damped Newton methods. Since these functions are not everywhere differentiable, Newton methods have been adapted to deal with B-differentiable functions. The main drawback of this approach is the need to globalize the results by means of a step-size procedure. We adapt a new method of projections on certain convex sets to solve the LCP. This approach becomes a Newton method with no need of stepsize. Both the theoretical and practical implications are encouraging. The convergence conditions extend with no modifications to a more general convex complementarity problem. If the procedure converges to a nondegenerate solution, the usual Newton quadratic rate of convergence is achieved.
- OSTI ID:
- 36050
- Report Number(s):
- CONF-9408161-; TRN: 94:009753-0319
- Resource Relation:
- Conference: 15. international symposium on mathematical programming, Ann Arbor, MI (United States), 15-19 Aug 1994; Other Information: PBD: 1994; Related Information: Is Part Of Mathematical programming: State of the art 1994; Birge, J.R.; Murty, K.G. [eds.]; PB: 312 p.
- Country of Publication:
- United States
- Language:
- English
Similar Records
The PATH solver: A stabilized Newton method for mixed complementarity problems
Computational experience in solving equilibrium models by a sequence of linear complementarity problems. Technical report SOL 83-1