CVM Optimization Toolkit

By Sayan Samanta and Axel van de Walle @ Brown University RI USA

Usage

usage: sro_correction [-h] [--seed SEED] [--disp] [--log LOG] [--out OUT] [--toscreen] [--eci ECI] [--vmat VMAT] [--clusters CLUSTERS]
                      [--maximal_clusters MAXIMAL_CLUSTERS] [--clustermult CLUSTERMULT] [--kikuchi_barker KIKUCHI_BARKER] [--configmult CONFIGMULT]
                      [--config CONFIG] [--lat LAT] [--Tmin TMIN] [--Tmax TMAX] [--Tstep TSTEP] [--inJoules] [--sro_method SRO_METHOD]
                      [--coeff_out COEFF_OUT] [--coeff_in COEFF_IN] [--fit_correction_only] [--norm_constraint] [--constr_tol CONSTR_TOL]
                      [--maxiter MAXITER] [--maxiter_linprog MAXITER_LINPROG] [--global_iterations GLOBAL_ITERATIONS] [--xtol XTOL] [--gtol GTOL]
                      [--barrier_tol BARRIER_TOL] [--initial_tr_radius INITIAL_TR_RADIUS] [--initial_constr_penalty INITIAL_CONSTR_PENALTY]
                      [--fit_ordered_only] [--method_linprog METHOD_LINPROG] [--basinhopping] [--verbose] [--approx_deriv] [--earlystop EARLYSTOP]
                      [--initial_stepsize INITIAL_STEPSIZE]

Parameters

Quick reference table

Short	Long	Default	Description
-h	--help		show this help message and exit
-s	--seed	42	set numpy random seed (for reproducible experiments)
-d	--disp	False	Flag to turn intermediate outputs on
-l	--log	log.out	Filename for the log file
-o	--out	result.csv	Indicates the name of the output file
-ts	--toscreen		Flag to output to screen in addition to log file
-e	--eci	eci.out	file containing ECI's
-vm	--vmat	vmat.out	file containing the vmatrix
-cl	--clusters	clusters.out	file containing cluster description
-mcl	--maximal_clusters	maxclus.in	file contain the maximal cluster coordinates
-clm	--clustermult	clusmult.out	file containing cluster multiplicities
-kb	--kikuchi_barker	configkb.out	file containing the kikuchi-barker coefficients
-com	--configmult	configmult.out	file containing cluster configuration multiplicities
-co	--config	config.out	file containing cluster configuration descriptions
-la	--lat	lat.in	contains the lattice description of the phase
-Tl	--Tmin	100	Minimum temperature for SRO correction
-Tm	--Tmax	2000	Maximum temperature for SRO correction
-Ts	--Tstep	100	Temperature increment for SRO correction
-ij	--inJoules		Sets energy units to Joules/mol
-sm	--sro_method	lm	method of fitting SRO correction model
-cout	--coeff_out	sro_coeffs.out	filename to store coefficients of SRO correction model
-cin	--coeff_in	sro_coeffs.in	filename containing initial coefficient for fitting SRO correction model
-fsr	--fit_correction_only		Flag to skip SRO correction optimisation but fit pre-existing data to SRO correction model
-nc	--norm_constraint		Flag to enable norm constraint
-ctol	--constr_tol	1e-08	Value for maximum allowed constraints violation
-mit	--maxiter	5000	Maximum no. of iterations for local minimizer
-mitlin	--maxiter_linprog	500000	Maximum no. of iterations for linear programming search
-git	--global_iterations	50	No. of global optimisations search steps
-x	--xtol	1e-08	Tolerance for termination of local minimizer by the change of correlations. The algorithm will terminate when tr_radius < xtol, where tr_radius is the radius of the trust region used in the algorithm
-g	--gtol	1e-08	The algorithm will terminate when both the infinity norm (i.e., max abs value) of the Lagrangian gradient and the constraint violation are smaller than gtol
-btol	--barrier_tol	1e-08	Threshold on the barrier parameter for the algorithm termination.
-itr	--initial_tr_radius	1	Initial trust radius. It reflects the trust the algorithm puts in the local approximation of the optimization problem. For an accurate local approximation the trust-region should be large and for an approximation valid only close to the current point it should be a small one.
-icp	--initial_constr_penalty	1	Initial Constraint Penalty. The penalty parameter is used for balancing the requirements of decreasing the objective function and satisfying the constraints.
-foo	--fit_ordered_only		Flag to find find_ordered state only and exit.
-linm	--method_linprog	highs	Method of linear programming for finding ordered correlations
-bh	--basinhopping	False	Flag to switch on basinhopping
-v	--verbose	0	Indicate the verbosity of the fit
-ad	--approx_deriv		Flag to enable estimation of derivatives
-es	--earlystop	20	Number of steps to break out of trials if no new minima has been found
-is	--initial_stepsize	0.1	Initial stepsize of the basinhopping algorithm from the disordered phase

-h, --help

show this help message and exit

--seed, -s (Default: 42)

set numpy random seed (for reproducible experiments)

--disp, -d (Default: False)

Flag to turn intermediate outputs on

--log, -l (Default: log.out)

Filename for the log file

--out, -o (Default: result.csv)

Indicates the name of the output file

--toscreen, -ts

Flag to output to screen in addition to log file

--eci, -e (Default: eci.out)

file containing ECI's

--vmat, -vm (Default: vmat.out)

file containing the vmatrix

--clusters, -cl (Default: clusters.out)

file containing cluster description

--maximal_clusters, -mcl (Default: maxclus.in)

file contain the maximal cluster coordinates

--clustermult, -clm (Default: clusmult.out)

file containing cluster multiplicities

--kikuchi_barker, -kb (Default: configkb.out)

file containing the kikuchi-barker coefficients

--configmult, -com (Default: configmult.out)

file containing cluster configuration multiplicities

--config, -co (Default: config.out)

file containing cluster configuration descriptions

--lat, -la (Default: lat.in)

contains the lattice description of the phase

--Tmin, -Tl (Default: 100)

Minimum temperature for SRO correction

--Tmax, -Tm (Default: 2000)

Maximum temperature for SRO correction

--Tstep, -Ts (Default: 100)

Temperature increment for SRO correction

--inJoules, -ij

Sets energy units to Joules/mol

--sro_method, -sm (Default: lm)

method of fitting SRO correction model

--coeff_out, -cout (Default: sro_coeffs.out)

filename to store coefficients of SRO correction model

--coeff_in, -cin (Default: sro_coeffs.in)

filename containing initial coefficient for fitting SRO correction model

--fit_correction_only, -fsr

Flag to skip SRO correction optimisation but fit pre-existing data to SRO correction model

--norm_constraint, -nc

Flag to enable norm constraint

--constr_tol, -ctol (Default: 1e-08)

Value for maximum allowed constraints violation

--maxiter, -mit (Default: 5000)

Maximum no. of iterations for local minimizer

--maxiter_linprog, -mitlin (Default: 500000)

Maximum no. of iterations for linear programming search

--global_iterations, -git (Default: 50)

No. of global optimisations search steps

--xtol, -x (Default: 1e-08)

Tolerance for termination of local minimizer by the change of correlations. The algorithm will terminate when tr_radius < xtol, where tr_radius is the radius of the trust region used in the algorithm

--gtol, -g (Default: 1e-08)

The algorithm will terminate when both the infinity norm (i.e., max abs value) of the Lagrangian gradient and the constraint violation are smaller than gtol

--barrier_tol, -btol (Default: 1e-08)

Threshold on the barrier parameter for the algorithm termination.

--initial_tr_radius, -itr (Default: 1)

Initial trust radius. It reflects the trust the algorithm puts in the local approximation of the optimization problem. For an accurate local approximation the trust-region should be large and for an approximation valid only close to the current point it should be a small one.

--initial_constr_penalty, -icp (Default: 1)

Initial Constraint Penalty. The penalty parameter is used for balancing the requirements of decreasing the objective function and satisfying the constraints.

--fit_ordered_only, -foo

Flag to find find ordered state only and exit.

--method_linprog, -linm (Default: highs)

Method of linear programming for finding ordered correlations

--basinhopping, -bh (Default: False)

Flag to switch on basinhopping

--verbose, -v (Default: 0)

Indicate the verbosity of the fit

--approx_deriv, -ad

Flag to enable estimation of derivatives

--earlystop, -es (Default: 20)

Number of steps to break out of trials if no new minima has been found

--initial_stepsize, -is (Default: 0.1)

Initial stepsize of the basinhopping algorithm from the disordered phase