EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
1419. Semi-Newton algorithms in nonsmooth difference programming
Invited abstract in session WB-42: Iterative Methods for Feasibility and Optimization Problems, stream Variational Analysis and Continuous Optimization.
Wednesday, 10:30-12:00Room: 98 (building: 306)
Authors (first author is the speaker)
| 1. | Francisco Javier Aragón Artacho
|
| Mathematics, University of Alicante |
Abstract
In this talk we present a new coderivative-based Newton algorithm for tackling unconstrained optimization problems whose objective function can be expressed as the difference of two generally nonconvex functions, namely, a continuously differentiable function with locally Lipschitzian derivatives and a locally Lipschitz and prox-regular function. Focusing our attention in the requirement of extended positive-definite properties of the generalized Hessian of the first function, we prove the well-posedness and various convergence results of the proposed algorithm. We conclude with some numerical experiments demonstrating the performance of the proposed algorithm in nonconvex settings.
This is a joint work with Boris S. Mordukhovich and Pedro Pérez-Aros.
Keywords
- Non-smooth Optimization
- Algorithms
- Continuous Optimization
Status: accepted
Back to the list of papers