FILTERSD uses a generalization of Robinson’s method, globalised by using a filter and trust region. The code makes use of a Ritz values approach Linear Constraint Problem solver. Second derivatives and storage of an approximate reduced Hessian matrix is avoided using a limited memory spectral gradient approach based on Ritz values. [LICENSE]
Bases: pyOpt.pyOpt_optimizer.Optimizer
FILTERSD Optimizer Class - Inherited from Optimizer Abstract Class
FILTERSD Optimizer Class Initialization
Keyword arguments:
Documentation last updated: Feb. 16, 2010 - Peter W. Jansen
Run Optimizer (Optimize Routine)
Keyword arguments:
Additional arguments and keyword arguments are passed to the objective function call.
Documentation last updated: February. 2, 2013 - Ruben E. Perez
Name | Type | Default Value | Notes |
---|---|---|---|
rho | float | 1.0 | Initial Trust Region Radius |
htol | float | 1.0e-6 | Constraint Feasibilities Tolerance |
rgtol | float | 1.0e-5 | Reduced Gradient l2 norm Tolerance |
maxit | int | 1000 | Maximum Number of Iterations |
maxgr | int | 1.0e+5 | Gradient Calls Upper Limit |
ubd | float | 1.0e+5 | Constraint Violation Upper Bound |
dchk | int | 0 | Derivative Check Flag (0 - no check, 1 - check) |
dtol | float | 1.0e-8 | Derivative Check Tolerance |
iprint | int | 1 | Print Flag (0-None, 1-Iter, 2-Debug) |
iout | int | 6 | Output Unit Number |
ifile | str | ‘FILTERSD.out’ | Output File Name |