Publication | Closed Access
Quadratic steepest descent on potential energy surfaces. III. Minima seeking along steepest descent lines
26
Citations
20
References
1993
Year
Mathematical ProgrammingNumerical AnalysisSearch OptimizationEngineeringQuadratic Steepest DescentVariational AnalysisQuadratic OptimizationUnconstrained OptimizationEnergy MinimizationTrue Steepest DescentQuasi-newton ProceduresDerivative-free OptimizationComputational GeometryLinear OptimizationMinimum SearchingPotential Energy SurfacesPhysicsInverse ProblemsComputer ScienceApproximation AlgorithmsNondifferentiable OptimizationNatural SciencesConvex OptimizationSteepest Descent LinesIterated Local Search
A simplified quadratic steepest descent method, based on the availability of energies and gradients, is formulated for use in minimum searching. It requires only a fraction of the computational effort needed for the previously developed accurate steepest descent procedures and, typically, involves less work than standard quasi-Newton minimum searches. At the same time, it follows the true steepest descent curves reasonably closely and reaches the closest minima. This is in contrast to quasi-Newton procedures which cannot be relied upon to do so. The performance is documented by applications to a variety of searches on the Müller–Brown surface.
| Year | Citations | |
|---|---|---|
Page 1
Page 1