A nonmonotone line search technique for Newton's method
SIAM Journal on Numerical Analysis
Practical methods of optimization; (2nd ed.)
Practical methods of optimization; (2nd ed.)
Convergence of quasi-Newton matrices generated by the symmetric rank one update
Mathematical Programming: Series A and B
A derivation of extrapolation algorithms based on error estimates
Proceedings of the 6th international congress on Computational and applied mathematics
Recent progress in unconstrained nonlinear optimization without derivatives
Mathematical Programming: Series A and B - Special issue: papers from ismp97, the 16th international symposium on mathematical programming, Lausanne EPFL
Testing Unconstrained Optimization Software
ACM Transactions on Mathematical Software (TOMS)
On the Global Convergence of Derivative-Free Methods for Unconstrained Optimization
SIAM Journal on Optimization
The Barzilai and Borwein Gradient Method for the Large Scale Unconstrained Minimization Problem
SIAM Journal on Optimization
Analysis of a Symmetric Rank-One Trust Region Method
SIAM Journal on Optimization
Introduction to Shape Optimization: Theory, Approximation, and Computation
Introduction to Shape Optimization: Theory, Approximation, and Computation
Journal of Computational and Applied Mathematics
Numerical Methods for Unconstrained Optimization and Nonlinear Equations (Classics in Applied Mathematics, 16)
Hi-index | 7.29 |
A tolerant derivative-free nonmonotone line-search technique is proposed and analyzed. Several consecutive increases in the objective function and also nondescent directions are admitted for unconstrained minimization. To exemplify the power of this new line search we describe a direct search algorithm in which the directions are chosen randomly. The convergence properties of this random method rely exclusively on the line-search technique. We present numerical experiments, to illustrate the advantages of using a derivative-free nonmonotone globalization strategy, with approximated-gradient type methods and also with the inverse SR1 update that could produce nondescent directions. In all cases we use a local variation finite differences approximation to the gradient.