BFGS Algorithm
- Command_Manual
- Tcl Commands
- Modeling_Commands
- model
- uniaxialMaterial
- ndMaterial
- frictionModel
- section
- geometricTransf
- element
- node
- sp commands
- mp commands
- timeSeries
- pattern
- mass
- block commands
- region
- rayleigh
- Analysis Commands
- Output Commands
- Misc Commands
- DataBase Commands
This command is used to construct a Broyden–Fletcher–Goldfarb–Shanno (BFGS) algorithm object. The BFGS method is one of the most effective matrix-update or quasi Newton methods for iteration on a nonlinear system of equations. The method computes new search directions at each iteration step based on the initial jacobian, and subsequent trial solutions. The unlike regular Newton-Raphson does not require the tangent matrix be reformulated and refactored at every iteration, however unlike ModifiedNewton it does not rely on the tangent matrix from a previous iteration.
algorithm BFGS |
REFERNCES:
- Denis, J.E "A Brief Survey of Convergence Methods for Quasi_Newton Methods", SIAMS-AMS Proceedings, Vol (9), 185-200, 1976.
- K.J. Bathe and A.P.Cimento "Some Practical Procedures for the Solution of Nonlinear Finte Element Equations", Computer Methods in Applied Mechanics and Engineering, Vol(22) 59-85, 1980.
THEORY:
In the BFGS method the following is employed at each time step:
- <math> \Gamma_{n+1} = R(U_{n+1}) - R(U_n),\!</math>
- <math> \Delta U_{n+1} = U_{n+1} + \Delta U\,\!</math>
Code Developed by: fmk