pass::problem Class Reference

#include <problem.hpp>

Inheritance diagram for pass::problem:

Public Member Functions
arma::vec	bounds_range () const noexcept
arma::uword	dimension () const noexcept
	problem (const arma::uword dimension, const double lower_bound, const double upper_bound, const std::string &name)
	problem (const arma::vec &lower_bounds, const arma::vec &upper_bounds, const std::string &name)
virtual double	evaluate (const arma::vec &agent) const =0
double	evaluate_normalised (const arma::vec &normalised_agent) const
arma::mat	normalised_random_agents (const arma::uword count) const
arma::mat	normalised_hammersley_agents (const arma::uword count) const
arma::mat	initialise_normalised_agents (const arma::uword count) const

Public Attributes
const arma::vec	lower_bounds
const arma::vec	upper_bounds
const std::string	name

Detailed Description

This class defines the interface for problems that can be approximated by a pass::optimiser. Subclasses need only implement evaluate.

Constructor & Destructor Documentation

problem() [1/2]

pass::problem::problem	(	const arma::uword	dimension,
		const double	lower_bound,
		const double	upper_bound,
		const std::string &	name
	)

Initialises an dimension-dimensional problem with uniform lower and upper bounds and a problem name.

problem() [2/2]

pass::problem::problem	(	const arma::vec &	lower_bounds,
		const arma::vec &	upper_bounds,
		const std::string &	name
	)

Initialises this problem with custom lower and upper bounds and a problem name. The problem dimension is derived from the bounds dimensions, which have to be equal.

Here is the call graph for this function:

Member Function Documentation

bounds_range()

arma::vec pass::problem::bounds_range ( ) const

noexcept

Returns the element-wise maximum difference between upper_bounds and lower_bounds.

Here is the caller graph for this function:

dimension()

arma::uword pass::problem::dimension ( ) const

noexcept

Returns the problem dimension.

Here is the caller graph for this function:

evaluate()

virtual double pass::problem::evaluate ( const arma::vec &  agent ) const

pure virtual

Evaluates this problem at agent, which must match the dimensions of this problem.

Implemented in pass::gtoc1, pass::rosenbrock_function, pass::rastrigin_function, pass::schwefel_function, pass::ackley_function, pass::styblinski_tang_function, pass::search_space_constraint, pass::de_jong_function, pass::griewank_function, pass::sum_of_different_powers_function, pass::evaluation_time_stall, pass::cassini1, pass::messenger_full, and pass::rosetta.

evaluate_normalised()

double pass::problem::evaluate_normalised

(

const arma::vec &

normalised_agent

)

const

Evaluates this problem at agent, which must be a normalized vector (all values must be in range [0, 1]). agent is mapped to the problem boundaries before evaluation.

Here is the caller graph for this function:

initialise_normalised_agents()

arma::mat pass::problem::initialise_normalised_agents

(

const arma::uword

count

)

const

Draws count distributed random agents from range [0, 1], stored column-wise. Agens are from type hammersley or random distributed (50%-50%)

Here is the caller graph for this function:

normalised_hammersley_agents()

arma::mat pass::problem::normalised_hammersley_agents

(

const arma::uword

count

)

const

Generates count hammersley points with the same dimension as this, stored column-wise.

For a definition of hammersley points, see http://www.cse.cuhk.edu.hk/~ttwong/papers/udpoint/udpoint.pdf, equations 1 to 3.

Here is the call graph for this function:

normalised_random_agents()

arma::mat pass::problem::normalised_random_agents

(

const arma::uword

count

)

const

Draws count uniformly distributed random agents from range [0, 1], stored column-wise.

Here is the caller graph for this function:

Member Data Documentation

lower_bounds

const arma::vec pass::problem::lower_bounds

Lower bound constraints for each problem dimension. An optimiser will never evaluate a point outside of these bounds.

Must be element-wise less than upper_bounds.

name

const std::string pass::problem::name

Identify every problem with its own name

upper_bounds

const arma::vec pass::problem::upper_bounds

Upper bound constraints for each problem dimension. An optimiser will never evaluate a point outside of these bounds.

Must be element-wise greater than upper_bounds.

---

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

• /Users/romeo/Desktop/PASS/include/pass_bits/problem.hpp
• /Users/romeo/Desktop/PASS/src/problem.cpp