Initial improvement of the hybrid accelerated gradient descent process

dc.contributor.author	Panić, Stefan
dc.contributor.author	Petrović, Milena
dc.contributor.author	Mihajlov Carević, Miroslava
dc.date.accessioned	2023-04-24T06:51:14Z
dc.date.available	2023-04-24T06:51:14Z
dc.date.issued	2018-08-01
dc.identifier.uri	https://platon.pr.ac.rs/handle/123456789/1250
dc.description.abstract	We improve the convergence properties of the iterative scheme for solving unconstrained optimisation problems introduced in Petrovic et al. [‘Hybridization of accelerated gradient descent method’, Numer. Algorithms (2017), doi:10.1007/s11075-017-0460-4] by optimising the value of the initial step length parameter in the backtracking line search procedure. We prove the validity of the algorithm and illustrate its advantages by numerical experiments and comparisons.	en_US
dc.language.iso	en_US	en_US
dc.publisher	Cambridge University Press	en_US
dc.title	Initial improvement of the hybrid accelerated gradient descent process	en_US
dc.title.alternative	Bulletin of the Australian Mathematical Society	en_US
dc.type	clanak-u-casopisu	en_US
dc.description.version	publishedVersion	en_US
dc.identifier.doi	https://doi.org/10.1017/S0004972718000552
dc.citation.volume	98
dc.citation.issue	2
dc.citation.spage	331
dc.citation.epage	338
dc.subject.keywords	line search	en_US
dc.subject.keywords	gradient descent methods	en_US
dc.subject.keywords	Newton method	en_US
dc.subject.keywords	convergence rate	en_US
dc.type.mCategory	M23	en_US
dc.type.mCategory	closedAccess	en_US
dc.type.mCategory	M23	en_US
dc.type.mCategory	closedAccess	en_US