SQP

Options table

Option	Type	Default Value	Description
problem	Problem or ProblemLite	None	Object containing the problem to be solved.
recording	bool	False	If True, record all outputs from the optimization. This needs to be enabled for hot-starting the same problem later, if the optimization is interrupted.
out_dir	str	None	The directory to store all the output files generated from the optimization.
hot_start_from	str	None	The record file from which to hot-start the optimization.
hot_start_atol	float	0.	The absolute tolerance check for the inputs when reusing outputs from the hot-start record.
hot_start_rtol	float	0.	The relative tolerance check for the inputs when reusing outputs from the hot-start record.
visualize	list	[]	The list of scalar variables to visualize during the optimization.
keep_viz_open	bool	False	If True, keeps the visualization window open after the optimization is complete.
turn_off_outputs	bool	False	If True, prevent modOpt from generating any output files.
maxiter	int	1000	Maximum number of major iterations.
opt_tol	float	1e-7	Optimality tolerance.
feas_tol	float	1e-7	Feasibility tolerance. Certifies convergence when the “scaled” maximum constraint violation is less than this value.
qp_tol	float	1e-4	Tolerance for the QP subproblem.
qp_maxiter	int	5000	Maximum number of iterations for the QP subproblem.
ls_min_step	float	1e-14	Minimum step size for the line search.
ls_max_step	float	1.	Maximum step size for the line search.
ls_maxiter	int	10	Maximum number of iterations for the line search.
ls_alpha_tol	float	1e-14	Relative tolerance for an acceptable step in the line search.
ls_eta_a	float	1e-4	Armijo (sufficient decrease condition) parameter for the line search.
ls_eta_w	float	0.9	Wolfe (curvature condition) parameter for the line search.
readable_outputs	list	[]	List of outputs to be written to readable text output files. Available outputs are: ‘major’, ‘obj’, ‘x’, ‘lag_mult’, ‘slacks’, ‘constraints’, ‘opt’, ‘feas’, ‘time’, ‘nfev’, ‘ngev’, ‘step’, ‘rho’, ‘merit’.

Source code

The source code is available here.

API

class modopt.SQP(problem, recording=False, out_dir=None, hot_start_from=None, hot_start_atol=0.0, hot_start_rtol=0.0, visualize=[], keep_viz_open=False, turn_off_outputs=False, **kwargs)

A Sequential Quadratic Programming (SQP) algorithm for general constrained optimization.

Parameters

problemProblem or ProblemLite

Object containing the problem to be solved.

recordingbool, default=False

If True, record all outputs from the optimization. This needs to be enabled for hot-starting the same problem later, if the optimization is interrupted.

out_dirstr, optional

The directory to store all the output files generated from the optimization.

hot_start_fromstr, optional

The record file from which to hot-start the optimization.

hot_start_atolfloat, default=0.

The absolute tolerance check for the inputs when reusing outputs from the hot-start record.

hot_start_rtolfloat, default=0.

The relative tolerance check for the inputs when reusing outputs from the hot-start record.

visualizelist, default=[]

The list of scalar variables to visualize during the optimization.

keep_viz_openbool, default=False

If True, keeps the visualization window open after the optimization is complete.

turn_off_outputsbool, default=False

If True, prevent modOpt from generating any output files.

maxiterint, default=1000

Maximum number of major iterations.

opt_tolfloat, default=1e-7

Optimality tolerance.

feas_tolfloat, default=1e-7

Feasibility tolerance. Certifies convergence when the “scaled” maximum constraint violation is less than this value.

qp_tolfloat, default=1e-4

Tolerance for the QP subproblem.

qp_maxiterint, default=5000

Maximum number of iterations for the QP subproblem.

ls_min_stepfloat, default=1e-14

Minimum step size for the line search.

ls_max_stepfloat, default=1.

Maximum step size for the line search.

ls_maxiterint, default=10

Maximum number of iterations for the line search.

ls_alpha_tolfloat, default=1e-14

Relative tolerance for an acceptable step in the line search.

ls_eta_afloat, default=1e-4

Armijo (sufficient decrease condition) parameter for the line search.

ls_eta_wfloat, default=0.9

Wolfe (curvature condition) parameter for the line search.

readable_outputslist, default=[]

List of outputs to be written to readable text output files. Available outputs are: ‘major’, ‘obj’, ‘x’, ‘lag_mult’, ‘slacks’, ‘constraints’, ‘opt’, ‘feas’, ‘time’, ‘nfev’, ‘ngev’, ‘step’, ‘rho’, ‘merit’.

Methods

check_first_derivatives([x, step, formulation])	Check the first derivatives of the optimization problem using finite differences.
print_results([summary_table, all])	Print the results of the optimization problem to the terminal.
solve()	Run the optimization algorithm to solve the problem.

check_first_derivatives(x=None, step=1e-06, formulation='rs')

Check the first derivatives of the optimization problem using finite differences.

Parameters

xnp.ndarray, optional

The design variables at which the derivatives are to be checked. If not provided, the initial design variables are used.

stepfloat, default=1e-6

The step size for the finite differences.

print_results(summary_table=False, all=False)

Print the results of the optimization problem to the terminal.

Parameters

summary_tablebool, default=False

If True, print the summary table for the optimization.

allbool, default=False

If False, print only the scalar outputs of the optimization problem. Otherwise, print all the outputs of the optimization problem.

solve()

Run the optimization algorithm to solve the problem.