En enhanced matrix-free method via double step length approach for solving systems of nonlinear equations
About this article
Abstract
A variant method for solving system of nonlinear equations is presented. This method use the special form of iteration with two step length parameters, we suggest a derivative-free method without computing the Jacobian via acceleration parameter as well as inexact line search procedure. The proposed method is proven to be globally convergent under mild condition. The preliminary numerical comparison reported in this paper using a large scale benchmark test problems show that the proposed method is practically quite effective.
Keywords
- Acceleration parameter
- Double step length
- Global Convergent.
References
[1] M.Y Waziri and Jamilu Sabiu, ” A derivative-free conjugate gradient method and its global convergence for symmetric nonlinear equations”, Journal of mathematics and mathematical sciences, vol.2015, pp.8.
[2] N.I Dbaruranovic-milicic, ” A multi step curve search algorithm in nonlinear optimization”, Yugoslav Journal of operation research, vol.18, ,2008 pp.47-52.
[3] N.Andrei, ” An accelerated gradient descent algorithm with backtracking for unconstrained optimization”, Numerical algorithm, vol.42, 2006, pp.63-73.
[4] L.Zang, W.Zhou and D.H Li, ” Global convergence of modified Fletcher-Reeves conjugate gradient method with Armijo-type line search”, Numerische mathematik, vol.164(1), 2005, pp.277-289.
[5] N.I Dbaruranovic-milicic and M. Gardasevic-Filipovic,” A multi step curve search algorithm in nonlinear convex case”, Facta universitatis series:mathematics and informatics, vol.25, 2010, pp.11-24.
[6] A.S Halilu, and M.YWaziri, ”A transformed double steplength method for for solving large-scale systems of nonlinear equations”, Journal of Numerical Mathematics and Stochastics, vol.9(1), 2017, pp.20-23.
[7] M.Raydan,” The Barzilai and Borwein gradient method for the large scale unconstrained minimization problem”, SIAM Journal of optimization, vol.7, 1997, pp.26-33.
[8] M.Raydan, ” On Barzilai and Borwein choice of step length for the gradient method”, IMA Journal of Numerical Analysis, vol.13, pp.321-326.
[9] D.Li and M.Fukushima, ” A global and superlinear convergent Gauss-Newton based BFGS method for symmetric nonlinear equation”, SIAM Journal on numerical Analysis, vol.37, 1999, pp.152-172.
[10] M.J petrovic, P.S Stanimirovic, ”Accelerated double direction method for solving unconstrained optimization problems”, Mathematical problem Eng.,2014.
[11] Gongli Yuan∗, Xiwen Lu,” A new backtracking inexact BFGS method for symmetric nonlinear equations”, Journal of computers and mathematics with applications, vol.55, 2008, pp.116-129.
[12] M.J petrovic, ”An accelerated double step size model in unconstrained optimization”, Journal of mathematics and computation, vol.250, 2015, pp.309-319.
[13] J.E Dennis, Jr and R.B Schnabel, ”Numerical method for unconstrained optimization and non-linear equations,practice Hall, Englewood cliffs, NJ, USA, 1983.
[14] E.Dolan and J.More ”Benchmarking optimization software with performance profiles”, Journal of Mathematical program, vol.91(2),2002, pp.201-213.