ConstrainedOptimizer

Overview¶

ConstrainedOptimizer is the abstract solver interface for constrained nonlinear optimization and tracks progress with dedicated parameter and state classes.

Authors: Frank Dellaert, et al. (see THANKS for the full author list)

See LICENSE for the license information

try:
    import google.colab
    %pip install --quiet gtsam-develop
except ImportError:
    pass

Key Concepts¶

ConstrainedOptimizerParams defines stopping tolerances, verbosity, and iteration limits.
ConstrainedOptimizerState tracks iteration number, values, cost, and constraint violations.
Derived optimizers implement optimize and typically solve a sequence of unconstrained subproblems.

Mathematical Formulation¶

ConstrainedOptimizer is the outer-loop interface for solving:

\min_x \frac{1}{2}\|f(x)\|^2\;\text{s.t.}\;h(x)=0,\;g(x)\le 0.

(1)

Concrete solvers repeatedly solve unconstrained subproblems and stop when cost and violation metrics meet absolute/relative tolerances.

Key User API¶

optimize() (implemented by derived solvers)
ConstrainedOptimizerParams: maxIterations, tolerance settings, verbosity
ConstrainedOptimizerState: values, cost, eqConstraintViolation, ineqConstraintViolation, violation()

Concise C++ Example¶

#include <gtsam/constrained/ConstrainedOptimizer.h>
#include <gtsam/constrained/PenaltyOptimizer.h>

using namespace gtsam;

auto params = std::make_shared<PenaltyOptimizerParams>();
params->maxIterations = 30;
params->absoluteViolationTolerance = 1e-6;

PenaltyOptimizer solver(problem, initialValues, params);
Values result = solver.optimize();