Enhancing structure relaxations for first-principles codes: an approximate Hessian approach

Simple item page
dc.creatorRondinelli, J. M.
dc.creatorDeng, B.
dc.creatorMarks, L. D.
dc.date2006-08-15
dc.date2006-11-13
dc.date.accessioned2026-07-07T08:48:41Z
dc.date.available2026-07-07T08:48:41Z
dc.descriptionWe present a method for improving the speed of geometry relaxation by using a harmonic approximation for the interaction potential between nearest neighbor atoms to construct an initial Hessian estimate. The model is quite robust, and yields approximately a 30% or better reduction in the number of calculations compared to an optimized diagonal initialization. Convergence with this initializer approaches the speed of a converged BFGS Hessian, therefore it is close to the best that can be achieved. Hessian preconditioning is discussed, and it is found that a compromise between an average condition number and a narrow distribution in eigenvalues produces the best optimization.
dc.description9 pages, 3 figures, added references, expanded optimization section
dc.identifierhttps://arxiv.org/abs/physics/0608160
dc.identifierhttp://arxiv.org/abs/physics/0608160
dc.identifierComput. Mater. Sci. 40 (2007) 345-353
dc.identifierdoi:10.1016/j.commatsci.2007.01.004
dc.identifier.urihttp://salesiana.dossiersoluciones.com/handle/123456789/144035
dc.subjectComputational Physics
dc.titleEnhancing structure relaxations for first-principles codes: an approximate Hessian approach
dc.typetext

Files

Collections

arXiv