Go to the documentation of this file.
29 typedef internal::NonlinearOptimizerState
State;
41 return linear->gradientAtZero();
72 const auto [newValues,
dummy] = nonlinearConjugateGradient<System, Values>(
Simple non-linear optimizer that solves using non-preconditioned CG.
Linear Factor Graph where all factors are Gaussians.
std::shared_ptr< This > shared_ptr
shared_ptr to this class
Private class for NonlinearOptimizer state.
DirectionMethod directionMethod_
std::shared_ptr< GaussianFactorGraph > linearize(const Values &linearizationPoint) const
Linearize a nonlinear factor graph.
NonlinearConjugateGradientOptimizer(const NonlinearFactorGraph &graph, const Values &initialValues, const Parameters ¶ms=Parameters(), const DirectionMethod &directionMethod=DirectionMethod::PolakRibiere)
Constructor.
static const SmartProjectionParams params
double error(const Values &values) const
Values retract(const VectorValues &delta) const
const NonlinearFactorGraph & graph_
internal::DoglegState State
def step(data, isam, result, truth, currPoseIndex, isamArgs=())
double error(const State &state) const
NonlinearFactorGraph graph_
The graph with nonlinear factors.
void g(const string &key, int i)
size_t iterations() const
return number of iterations in current optimizer state
static VectorValues gradientInPlace(const NonlinearFactorGraph &nfg, const Values &values)
Return the gradient vector of a nonlinear factor graph.
const Values & optimize() override
State advance(const State ¤t, const double alpha, const Gradient &g) const
NonlinearFactorGraph graph
GaussianFactorGraph::shared_ptr iterate() override
std::tuple< V, int > nonlinearConjugateGradient(const S &system, const V &initial, const NonlinearOptimizerParams ¶ms, const bool singleIteration, const DirectionMethod &directionMethod=DirectionMethod::PolakRibiere, const bool gradientDescent=false)
std::unique_ptr< internal::NonlinearOptimizerState > state_
PIMPL'd state.
Gradient gradient(const State &state) const
A non-templated config holding any types of Manifold-group elements.
gtsam
Author(s):
autogenerated on Sun Dec 22 2024 04:12:23