Multi-steps Symmetric Rank-one Update for Unconstrained Optimization
نویسندگان
چکیده
In this paper, we present a generalized Symmetric Rank-one (SR1) method by employing interpolatory polynomials in order to possess a more accurate information from more than one previous step. The basic idea is to incorporate the SR1 update within the framework of multi-step methods. Hence iterates could be interpolated by a curve in such a way that the consecutive points define the curves. However to preserve the positive definiteness of the SR1 updates a restart procedure is applied, in which we restart the SR1 update by a scale of the identity. Comparison to multi-steps BFGS method, the proposed algorithm shows significant improvements in numerical results. Mathematics subject classification: 65K10 • 90C53
منابع مشابه
Memoryless Modified Symmetric Rank-One Method for Large-Scale Unconstrained Optimization
Problem statement: Memoryless QN methods have been regarded effective techniques for solving large-scale problems that can be considered as one step limited memory QN methods. In this study, we present a scaled memoryless modified Symmetric Rank-One (SR1) algorithm and investigate the numerical performance of the proposed algorithm for solving large-scale unconstrained optimization problems. Ap...
متن کاملA new quasi-Newton pattern search method based on symmetric rank-one update for unconstrained optimization
This paper proposes a new robust and quickly convergent pattern search method based on an implementation of OCSSR1 (Optimal Conditioning Based Self-Scaling Symmetric Rank-One) algorithm [M.R. Osborne, L.P. Sun, A new approach to symmetric rank-one updating, IMA Journal of Numerical Analysis 19 (1999) 497–507] for unconstrained optimization. This method utilizes the factorization of approximatin...
متن کاملStructured symmetric rank-one method for unconstrained optimization
In this paper, we investigate a symmetric rank-one (SR1) quasi-Newton (QN) formula in which the Hessian of the objective function has some special structure. Instead of approximating the whole Hessian via the SR1 formula, we consider an approach which only approximates part of the Hessian matrix that is not easily acquired. Although the SR1 update possesses desirable features, it is unstable in...
متن کاملA symmetric rank-one quasi-Newton line-search method using negative curvature directions
We propose a quasi-Newton line-search method that uses negative curvature directions for solving unconstrained optimization problems. In this method, the symmetric rank-one (SR1) rule is used to update the Hessian approximation. The SR1 update rule is known to have a good numerical performance; however, it does not guarantee positive definiteness of the updated matrix. We first discuss the deta...
متن کاملConvergence of quasi-Newton matrices generated by the symmetric rank one update
Quasi-Newton algorithms for unconstrained nonlinear minimization generate a sequence of matrices that can be considered as approximations of the objective function second derivatives. This paper gives conditions under which these approximations can be proved to converge globally to the true Hessian matrix, in the case where the Symmetric Rank One update formula is used. The rate of convergence ...
متن کامل