var_data::var_control_type Type Reference

Public Attributes
character(6)	opt_method
	which optimization method to use currently this has a number of options: More...
integer	cg_method
	which type of nonlinear CG method to use options are: 1 FLETCHER-REEVES 2 POLAK-RIBIERE (DEFAULT) 3 POSITIVE POLAK-RIBIERE (BETA=MAX{BETA,0} ) More...
real(kind=kind(1.0d0))	cg_eps
	convergence tolerance for CG method DEFAULT = 1.0d-5 More...
real(kind=kind(1.0d0))	lbfgs_factr
	factr is a DOUBLE PRECISION variable that must be set by the user. It is a tolerance in the termination test for the algorithm. The iteration will stop when More...
real(kind=kind(1.0d0))	lbfgs_pgtol
	pgtol is a double precision variable. On entry pgtol >= 0 is specified by the user. The iteration will stop when More...
real(kind=kind(1.0d0)), dimension(:), allocatable	l
real(kind=kind(1.0d0)), dimension(:), allocatable	u
real(kind=kind(1.0d0)), dimension(:), allocatable	x0
integer, dimension(:), allocatable	nbd
integer	n
	the size of the state vector More...
integer	total_timesteps
	the total number of timesteps in the assimilation window More...
integer, dimension(:), allocatable	ny
	array containing the number of observations. More...

Detailed Description

Definition at line 32 of file var_data.f90.

Member Data Documentation

real(kind=kind(1.0d0)) var_data::var_control_type::cg_eps

convergence tolerance for CG method
DEFAULT = 1.0d-5

Definition at line 44 of file var_data.f90.

integer var_data::var_control_type::cg_method

which type of nonlinear CG method to use options are:
1 FLETCHER-REEVES
2 POLAK-RIBIERE (DEFAULT)
3 POSITIVE POLAK-RIBIERE (BETA=MAX{BETA,0} )

Definition at line 39 of file var_data.f90.

real(kind=kind(1.0d0)), dimension(:), allocatable var_data::var_control_type::l

Definition at line 78 of file var_data.f90.

real(kind=kind(1.0d0)) var_data::var_control_type::lbfgs_factr

factr is a DOUBLE PRECISION variable that must be set by the user.
It is a tolerance in the termination test for the algorithm. The iteration will stop when

(f^k - f^{k+1})/max{|f^k|,|f^{k+1}|,1} <= factr*epsmch

where epsmch is the machine precision which is automatically generated by the code. Typical values for factr on a computer with 15 digits of accuracy in double precision are:
factr=1.d+12 for low accuracy;
1.d+7 for moderate accuracy;
1.d+1 for extremely high accuracy.
The user can suppress this termination test by setting factr=0.

DEFAULT = 1.0d7

Definition at line 48 of file var_data.f90.

real(kind=kind(1.0d0)) var_data::var_control_type::lbfgs_pgtol

pgtol is a double precision variable.
On entry pgtol >= 0 is specified by the user. The iteration will stop when

max{|proj g_i | i = 1, ..., n} <= pgtol

where pg_i is the ith component of the projected gradient.
The user can suppress this termination test by setting pgtol=0.

DEFAULT = 1.0d-5

Definition at line 66 of file var_data.f90.

integer var_data::var_control_type::n

the size of the state vector

Definition at line 81 of file var_data.f90.

integer, dimension(:), allocatable var_data::var_control_type::nbd

Definition at line 79 of file var_data.f90.

integer, dimension(:), allocatable var_data::var_control_type::ny

array containing the number of observations.

ny(t) contains the number of observations at time t

if no observations at time t then ny(t) = 0

Definition at line 85 of file var_data.f90.

character(6) var_data::var_control_type::opt_method

which optimization method to use currently this has a number of options:

• 'cg'
  
• 'lbfgs'
  
• 'lbfgsb'
  

Definition at line 33 of file var_data.f90.

integer var_data::var_control_type::total_timesteps

the total number of timesteps in the assimilation window

Definition at line 82 of file var_data.f90.

real(kind=kind(1.0d0)), dimension(:), allocatable var_data::var_control_type::u

Definition at line 78 of file var_data.f90.

real(kind=kind(1.0d0)), dimension(:), allocatable var_data::var_control_type::x0

Definition at line 78 of file var_data.f90.

---

The documentation for this type was generated from the following file:

• src/4dEnVar/var_data.f90