/IMPL/LBFGS - Parameters of BFGS quasi-Newton Method for Implicit Nonlinear
Description
Change the maximum number of BFGS quasi-Newton method for implicit nonlinear.
Format
/IMPL/LBFGS/L
L
Data |
Description |
L |
Maximum number of BFGS vectors |
1. | /IMPL/NONLIN/2 should be defined. |
2. | If L =0 or /IMPL/LBFGS/L is not defined, L = L_A is defined in /IMPL/NONLIN/2. |
When this keyword is defined, the new update vectors are accumulated over the old ones by sliding down the set of vectors loosing the oldest ones.
See Also: