pass Namespace Reference

Namespaces
	gtoc

Classes
class	ackley_function
struct	asteroid
class	cassini1
struct	celestial_body
class	de_jong_function
class	evaluation_time_stall
class	griewank_function
class	gtoc1
struct	lambert_solution
class	messenger_full
struct	optimise_result
class	optimiser
class	parallel_swarm_search
class	particle_swarm_optimisation
class	problem
class	random_search
class	rastrigin_function
class	regression
class	rosenbrock_function
class	rosetta
class	schwefel_function
class	search_space_constraint
class	seed
class	stopwatch
class	styblinski_tang_function
class	sum_of_different_powers_function

Typedefs
typedef struct pass::lambert_solution	lambert_solution

Functions
void	search_parameters (const pass::problem &problem, const bool benchmark)
arma::mat	parameter_evaluate (pass::optimiser &optimiser, const pass::problem &problems)
arma::uword	compare_segments (const arma::mat first_segment_runtimes, const arma::mat second_segment_runtimes)
bool	enable_openmp (const pass::problem &problem)
arma::mat	train (const int &examples)
arma::rowvec	build_model (const arma::mat &training_points)
double	problem_evaluation_time (const pass::problem &problem)
int	thread_number ()
int	number_of_threads ()
int	node_rank ()
int	number_of_nodes ()
lambert_solution	lambert (std::array< double, 3 > r1_in, std::array< double, 3 > r2_in, double t, const double mu, const int lw)
std::pair< double, double >	pow_swing_by_inv (const double Vin, const double Vout, const double alpha)
std::pair< std::array< double, 3 >, std::array< double, 3 > >	conversion (const std::array< double, 6 > &E, const double mu)
double	mean_to_eccentric (const double m, const double e)
void	pow_swing_by_inv (const double, const double, const double, double &, double &)
double	random_double_uniform_in_range (double min, double max)
int	random_integer_uniform_in_range (int min, int max)
arma::rowvec	integers_uniform_in_range (const int min, const int max, const int count)
arma::vec	random_neighbour (const arma::vec &agent, const double minimal_distance, const double maximal_distance)
bool	is_verbose (false)
int	global_number_of_runs (1)
double	precision (1e-06)
arma::uword	parameter_setting_number_of_runs (10)

Variables
bool	is_verbose
int	global_number_of_runs
double	precision
arma::uword	parameter_setting_number_of_runs
const std::array< arma::uword, 2000 >	prime_numbers

Typedef Documentation

lambert_solution

typedef struct pass::lambert_solution pass::lambert_solution

Function Documentation

build_model()

arma::rowvec pass::build_model

(

const arma::mat &

training_points

)

Machine Learning - Step 2.

Builds the model from the trainingdata. Tests two models: linear and polynomial.

Here is the call graph for this function:

Here is the caller graph for this function:

compare_segments()

arma::uword pass::compare_segments	(	const arma::mat	first_segment_runtimes,
		const arma::mat	second_segment_runtimes
	)

Compare the two segments with each other.

Returns

• '1' if the first segmet is the best
• '2' if the second segmet is the best

Here is the caller graph for this function:

conversion()

std::pair< std::array< double, 3 >, std::array< double, 3 > > pass::conversion	(	const std::array< double, 6 > &	E,
		const double	mu
	)

This function is an adapted version of function Conversion in file ./astro_functions.hpp.

Returns a tuple (position, velocity).

Here is the caller graph for this function:

enable_openmp()

bool pass::enable_openmp

(

const pass::problem &

problem

)

Depending on your problem it returns if you should use openMP or not Outputs are given through all the analyse process.

Steps:

1. Estimate the evaluation time of your problem
2. Generate trainingsdata
3. Test linear model if it fits
4. Test polynomial model if it fitst
5. Predict the speedup for your model
6. Give suggestions if to activate openMP or not

Here is the call graph for this function:

global_number_of_runs()

int pass::global_number_of_runs

(

1

)

Global variables used for evaluations : to help saving raw data Starts from '1'

integers_uniform_in_range()

arma::rowvec pass::integers_uniform_in_range	(	const int	min,
		const int	max,
		const int	count
	)

Returns a uniformly 'count' drawn random integer numbers in range [min, max].

Here is the caller graph for this function:

is_verbose()

bool pass::is_verbose

(

false

)

Global variables used for evaluations : analyse the behaviour of the algorithm Is initialized to false.

lambert()

pass::lambert_solution pass::lambert	(	std::array< double, 3 >	r1_in,
		std::array< double, 3 >	r2_in,
		double	t,
		const double	mu,
		const int	lw
	)

This function is an adapted version of function LambertI in file ./lambert.h. For more details and documentation, please take a look at the original file.

Throws std::invalid_argument if t is negative or zero.

Here is the call graph for this function:

Here is the caller graph for this function:

mean_to_eccentric()

double pass::mean_to_eccentric	(	const double	m,
		const double	e
	)

This function is an adapted version of function Mean2Eccentric in file ./astro_functions.h.

Here is the call graph for this function:

Here is the caller graph for this function:

node_rank()

int pass::node_rank

(

)

Use MPI : returns which number a.k.a rank the node have

Here is the caller graph for this function:

number_of_nodes()

int pass::number_of_nodes

(

)

Use MPI : returns the number of the nodes

Here is the caller graph for this function:

number_of_threads()

int pass::number_of_threads

(

)

Use OpenMP : returns the number of the threads

Here is the caller graph for this function:

parameter_evaluate()

arma::mat pass::parameter_evaluate	(	pass::optimiser &	optimiser,
		const pass::problem &	problems
	)

Evaluate the Problem.

The Problem will be evaluated 10 times. Set the global variable to change that pass::parameter_setting_number_of_runs

Here is the call graph for this function:

Here is the caller graph for this function:

parameter_setting_number_of_runs()

arma::uword pass::parameter_setting_number_of_runs

(

10

)

Global variables used for evaluations : number of the runs Is initialized to 10.

pow_swing_by_inv() [1/2]

void pass::pow_swing_by_inv	(	const double	Vin,
		const double	Vout,
		const double	alpha,
		double &	DV,
		double &	rp
	)

pow_swing_by_inv() [2/2]

std::pair< double, double > pass::pow_swing_by_inv	(	const double	Vin,
		const double	Vout,
		const double	alpha
	)

This function is an adapted version of function pow_swing_by_inv in file ./pow_swing_by_inv.hpp.

Returns a tuple (DV, rp).

Here is the caller graph for this function:

precision()

double pass::precision

(

1e-

06

)

Global variables used for evaluations : the precision to stop algorithms and set restart Is initialized to 1e-06.

problem_evaluation_time()

double pass::problem_evaluation_time

(

const pass::problem &

problem

)

Returns the evaluation time of a problem in nanoseconds.

NOTE: It is not a requirement to get the same output always because the CPU can be less or more used by other processes running on the computer. In the human mind, we can remember the solution of a math problem, though for a computer the same process will always be something new, so, it is not required to get the same result always!

Here is the call graph for this function:

Here is the caller graph for this function:

random_double_uniform_in_range()

double pass::random_double_uniform_in_range	(	double	min,
		double	max
	)

Returns a uniformly drawn random double number in range [min, max].

Here is the caller graph for this function:

random_integer_uniform_in_range()

int pass::random_integer_uniform_in_range	(	int	min,
		int	max
	)

Returns a uniformly drawn random integer number in range [min, max].

Here is the caller graph for this function:

random_neighbour()

arma::vec pass::random_neighbour	(	const arma::vec &	agent,
		const double	minimal_distance,
		const double	maximal_distance
	)

Return a randomly and uniformly drawn neighbour of agent, with distance [minimal_distance, maximal_distance].

Here is the call graph for this function:

Here is the caller graph for this function:

search_parameters()

void pass::search_parameters	(	const pass::problem &	problem,
		const bool	benchmark
	)

Search the best parameters for the parallel swarm search algorithm and saves the parameters in a file which can be than loaded.

Following parameter of the parallel swarm search are investigated:

• Swarm Size
• Neighbourhood Probability
• Inertia
• Cognitive Acceleration = Social Acceleration

• problem: the given problem
• Benchmark? 'True' or 'False'

Here is the call graph for this function:

thread_number()

int pass::thread_number

(

)

Use OpenMP : returns which number the thread have

train()

arma::mat pass::train

(

const int &

examples

)

Machine Learning - Step 1.

Collects trainingdata for the model. Default: 120 trainingspoint are collected The more data you collect (better data), the better the model performs.

Here is the call graph for this function:

Here is the caller graph for this function:

Variable Documentation

global_number_of_runs

int pass::global_number_of_runs

Global variables used for evaluations : to help saving raw data Starts from '1'

is_verbose

bool pass::is_verbose

Global variables used for evaluations : analyse the behaviour of the algorithm Is initialized to false.

parameter_setting_number_of_runs

arma::uword pass::parameter_setting_number_of_runs

Global variables used for evaluations : number of the runs Is initialized to 10.

precision

double pass::precision

Global variables used for evaluations : the precision to stop algorithms and set restart Is initialized to 1e-06.

prime_numbers

const std::array< arma::uword, 2000 > pass::prime_numbers

Primes for use in pass::problem::hammersley_agents.

Taken from https://cis.temple.edu/~beigel/cis573/alizzy/prime-list.html