autoware::common::optimization Namespace Reference

Namespaces
	detail

Classes
class	CachedExpression
class	CacheStateMachine
class	ComputeMode
class	EigenComparator
	Generic equality comparison functor for eigen matrices. More...
class	Expression
class	FixedLineSearch
	Class to use a fixed step length during optimization. More...
class	LineSearch
	Base class (CRTP) to mange the step length during optimization. More...
class	MoreThuenteLineSearch
	This class describes a More-Thuente line search as presented in the paper "Line Search Algorithms with Guaranteed Sufficient Decrease" by Jorge J. More and David J. Thuente. More...
class	NewtonsMethodOptimizer
	Optimizer using the Newton method with line search. More...
class	OptimizationOptions
class	OptimizationProblem< Derived, DomainValueT, NumJacobianColsT, NumVarsT, ObjectiveT, std::tuple< EqualityConstraints... >, std::tuple< InequalityConstraints... > >
class	OptimizationSummary
class	Optimizer
class	UnconstrainedOptimizationProblem

Enumerations
enum	TerminationType { TerminationType::CONVERGENCE = 0U, TerminationType::NO_CONVERGENCE = 1U, TerminationType::FAILURE = 2U }
	Type of termination at the end of an optimization loop. More...
enum	ExpressionTerm { ExpressionTerm::SCORE, ExpressionTerm::JACOBIAN, ExpressionTerm::HESSIAN }
	Enum class representing expression terms. More...

Variables
template<typename Derived , typename DomainValueT , int NumJacobianColsT, int NumVarsT, typename ObjectiveT , typename EqualityConstraintsT , typename InequalityConstraintsT >
class OPTIMIZATION_PUBLIC	OptimizationProblem

Enumeration Type Documentation

ExpressionTerm

enum autoware::common::optimization::ExpressionTerm

strong

Enum class representing expression terms.

Enumerator
SCORE
JACOBIAN
HESSIAN

TerminationType

enum autoware::common::optimization::TerminationType

strong

Type of termination at the end of an optimization loop.

Enumerator
CONVERGENCE
NO_CONVERGENCE
FAILURE

Variable Documentation

OptimizationProblem

template<typename Derived , typename DomainValueT , int NumJacobianColsT, int NumVarsT, typename ObjectiveT , typename EqualityConstraintsT , typename InequalityConstraintsT >

class OPTIMIZATION_PUBLIC autoware::common::optimization::OptimizationProblem

A generalized representation of an optimization problem. Constrained implementation classes must implement the following methods: score_opt_(x); jacobian_opt_(x, out); hessian_opt_(x, out);

Template Parameters

Derived	Implementation class
DomainValueT	The type for parameter to be optimized for.
NumJacobianColsT	Number of columns in the jacobian matrix.
NumVarsT	Number of variables.
ObjectiveT	Objective implementation class.
EqualityConstraintsT	A tuple of equality constraint expressions.
InequalityConstraintsT	A tuple of inequality constraint expressions.