mc::optim::GA Class Reference

Genetic Algorithm optimizer. More...

#include <GA.hpp>

Inheritance diagram for mc::optim::GA:

Collaboration diagram for mc::optim::GA:

Classes
struct	Individual
	A single population member. More...

Public Member Functions
	GA (const GAConfig &config=GAConfig{})
	Construct a GA optimizer with the given configuration.
void	setObjectiveFunction (ObjectiveFunction func) override
	Set the objective function to optimize.
void	setBounds (const Coordinates &lower, const Coordinates &upper) override
	Set lower/upper bounds of the search hyper-rectangle.
void	setMode (OptimizationMode mode) override
	Set optimization mode (minimize or maximize).
void	setCallback (StepCallback cb) override
	Register a callback invoked after each generation.
Solution	optimize () override
	Run GA for max_generations.
void	step () override
	Perform one generation: selection, crossover, mutation, evaluation, and serial update of the global best.
Solution	getBestSolution () const override
	Get the best solution found so far.
const std::vector< Individual > &	getPopulation () const
	Access the current population.
Public Member Functions inherited from mc::optim::Optimizer
virtual	~Optimizer ()=default

Additional Inherited Members
Public Types inherited from mc::optim::Optimizer
using	StepCallback = std::function< void(const Solution &current_best, size_t iteration)>
	Callback invoked after each step/generation.

Detailed Description

Genetic Algorithm optimizer.

Note

Determinism: selection/variation use a shared RNG; fitness evaluation can be parallelized. Best-solution tracking is updated serially.

Definition at line 46 of file GA.hpp.

Constructor & Destructor Documentation

GA()

mc::optim::GA::GA ( const GAConfig &  config = GAConfig{} )

explicit

Construct a GA optimizer with the given configuration.

Definition at line 19 of file GA.cpp.

Member Function Documentation

getBestSolution()

Solution mc::optim::GA::getBestSolution ( ) const

overridevirtual

Get the best solution found so far.

Implements mc::optim::Optimizer.

Definition at line 224 of file GA.cpp.

getPopulation()

const std::vector< Individual > & mc::optim::GA::getPopulation ( ) const

inline

Access the current population.

Definition at line 101 of file GA.hpp.

optimize()

Solution mc::optim::GA::optimize ( )

overridevirtual

Run GA for max_generations.

Returns

The best solution found.

Implements mc::optim::Optimizer.

Definition at line 211 of file GA.cpp.

Here is the call graph for this function:

setBounds()

void mc::optim::GA::setBounds ( const Coordinates &  lower, const Coordinates &  upper  )

overridevirtual

Set lower/upper bounds of the search hyper-rectangle.

Exceptions

std::invalid_argument

if dimensions mismatch.

Implements mc::optim::Optimizer.

Definition at line 34 of file GA.cpp.

setCallback()

void mc::optim::GA::setCallback ( StepCallback  cb )

overridevirtual

Register a callback invoked after each generation.

Parameters

cb	Receives current global-best and generation index.

Implements mc::optim::Optimizer.

Definition at line 30 of file GA.cpp.

setMode()

void mc::optim::GA::setMode ( OptimizationMode  mode )

overridevirtual

Set optimization mode (minimize or maximize).

Implements mc::optim::Optimizer.

Definition at line 43 of file GA.cpp.

Here is the call graph for this function:

setObjectiveFunction()

void mc::optim::GA::setObjectiveFunction ( ObjectiveFunction  func )

overridevirtual

Set the objective function to optimize.

Parameters

func	Function mapping coordinates to scalar cost/fitness.

Implements mc::optim::Optimizer.

Definition at line 25 of file GA.cpp.

step()

void mc::optim::GA::step ( )

overridevirtual

Perform one generation: selection, crossover, mutation, evaluation, and serial update of the global best.

Implements mc::optim::Optimizer.

Definition at line 150 of file GA.cpp.

---

The documentation for this class was generated from the following files:

• src/montecarlo/optimizers/GA.hpp
• src/montecarlo/optimizers/GA.cpp