Config

class moba_muscod.muscod_config.Config

Object oriented representation of MUSCOD’s configuration (.dat) file.

acc

Termination accuracy. Maximum acceptable KKT tolerance.

ftol

Final integration tolerance. Should be lower than termination accuracy.

itmax

Maximum number of SQP iterations.

itol

Initial integration tolerance.

levmar

Levenberg-Maquardt regularization factor for the Hessian.

libcond

Condensing algorithm for the condensing of the block sparse QP.

Possible choices are:

• ‘cond_std’: Standard condensing algorithm. Condenses state and control discretization (default).
• ‘cond_min’: Minimal condensing algorithm. Only condenses control discretization. Usefull for sparse QP solvers.

libhessian

Hessian approximation method.

Possible choices are:

• ‘hess_update’: BFGS update (default)
• ‘hess_finitediff’: exact computation with finite differences
• ‘hess_const’: constant Hessian
• ‘hess_gaussnewton’: Gauss Newton Approximation (only for LSQ)
• ‘hess_gaussnewton_up’
• ‘hess_limitedmemoryupdate’: limited memory BFGS update

libmodel

Name of compiled model (.dll)

libmssqp

Multiple shooting SQP algorithm.

libqps

QP (Quadratic Programming) solver module.

libsolve

Globalization strategy for the SQP algorithm.

Possible choices are:

• ‘solve_slse’: Standard line search (default)
• ‘solve_fullstep’: Disable globalization and always perform the full SQP step
• ‘solve_tbox’: Boxstep trust region technique.
• ‘solve_alf’: Schittkowski augmented line search
• ‘solve_vmcwd’: Line search with a watchdog technique.
• ‘solve_wdog’: Watchdog technique using Powell’s penalty function.

libtchk

Termination check module for the SQP algorithm.

nhtopy

Allows to employ a homotopy strategy.

Beginning with a lower initial integration tolerance (and thus faster SQP iterations) set via options_itol, the current integration tolerance ctol is tightened until – after nhtopy steps – ftol (set via options_ftol) is attained. The steps in ctol interpolate logarithmically between itol and ftol. A homotopy step is performed during the SQP iterations whenever the current accuracy cacc is attained. It is determined as cacc= ctol * acc / ftol.

np

Number of global parameters.

of_scale

Objective scale.

plbound

Lower bounds of global parameters.

plevel_screen

Print level for printout to the screen or console.

pname

Names of global parameters.

pscale

Scale values of global parameters.

pstart

Start values of global parameters.

pubound

Upper bounds of global parameters.

qp_featol

Feasibility tolerance of the QP solver.

qp_itmax

Maximum number of iteration of the QP solver per SQP iteration.

qp_relax

Constraint relaxation factor for infeasible QPs.

When an infeasible QP is detected, all infeasible constraints are shifted by qp_relax times the amount of infeasibility, to make the next QP guaranteedly feasible. Useful values are greater than 1.0.

rfac

Regularization factor.

To make some non unique problems solvable. If the original problem has a flat minimum this option may be useful, as it adds a tiny quadratic term of all problem variables to the objective function to define the minimum uniquely.

s_spec

Specification mode for state variable start values.