The Journal of Chemical Physics · 1993 · 58 citations · 8 references
Natural SciencesChemistryComputational ChemistryOptimizationSurface ScienceEnergy MinimizationGradient-based OptimizationNumerical AnalysisPotential Energy SurfacesSurface EnergeticsAnalytic Potential Energy SurfacesHessian ApproximationPotential Energy Surface AnalysisPotential Energy Surface OptimizationSaddle Point SearchSaddle PointsSecond-order OptimizationSteepest Descent
A second order method is developed for determining steepest descent lines of potential energy surfaces by following steepest curves of successive local quadratic surface approximations. The basic principle is similar to that of a previously developed method where, however, the availability of analytically calculated exact Hessians was assumed wherever needed. By contrast, only the analytically calculated exact values of the energy and its gradient are used here and this difference entails marked changes in strategy. Applications to the Gonzalez–Schlegel and the Müller–Brown surfaces show that the method compares favorably with existing methods.
8
A new approach to variable metric algorithms
R. Fletcher · The Computer Journal · 1970 · 4.1K citations · Full text
Conditioning of quasi-Newton methods for function minimization
David F. Shanno · Mathematics of Computation · 1970 · 3.7K citations · Full text