Population package

This package contains the building blocks that create the particle and the swarm objects.

Submodules

star_pso.population.jat_particle module

class star_pso.population.jat_particle.JatParticle(container: list[DataBlock], value: float = -inf, best_value: float = -inf)[source]

Bases: object

Description:

Implements a dataclass for the ‘Jack of all trades’ particle. This dataclass maintains a container (list of data blocks) and the rest of the information is held in its data blocks.

property best_position: list

Accessor of the particle’s best position.

Returns:

a list with each data block best position.

property best_value: float

Accessor of the best function value.

Returns:

(float) best function value.

property container: list[DataBlock]

Accessor of the container list of the particle.

Returns:

the list (of data blocks) of the particle.

property position: list

Accessor of the particle’s position.

Returns:

a list with each data block position.

reset_position() → None[source]

Iterate through all the data blocks in the container and reset their positions.

Returns:

None.

property size: int

Returns the size (or length) of the particle.

Since the size of the container doesn’t change dynamically we can cache this value for faster retrieval.

Returns:

(int) the length of the particle.

property value: float

Accessor of the current function value.

Returns:

(float) function value.

star_pso.population.particle module

class star_pso.population.particle.Particle(initial_position: _SupportsArray[dtype[Any]] | _NestedSequence[_SupportsArray[dtype[Any]]] | bool | int | float | complex | str | bytes | _NestedSequence[bool | int | float | complex | str | bytes])[source]

Bases: object

Description:

Models the particle in the swarm of the PSO.

property best_position: _SupportsArray[dtype[Any]] | _NestedSequence[_SupportsArray[dtype[Any]]] | bool | int | float | complex | str | bytes | _NestedSequence[bool | int | float | complex | str | bytes]

Returns the best (so far) position of the particle object.

Returns:

(ArrayLike) position vector.

property best_value: float

Accessor of the best function value recorded.

Returns:

(float) best function value.

property position: _SupportsArray[dtype[Any]] | _NestedSequence[_SupportsArray[dtype[Any]]] | bool | int | float | complex | str | bytes | _NestedSequence[bool | int | float | complex | str | bytes]

Accessor of the positions in the particle.

Returns:

the ArrayLike (vector) of the positions.

property size: int

Returns the size (or length) of the particle.

Returns:

(int) the length of the particle.

property value: float

Accessor of the current function value.

Returns:

(float) function value.

star_pso.population.swarm module

class star_pso.population.swarm.Swarm(population: list[Particle | JatParticle], copy: bool = False)[source]

Bases: object

Description:

Implements a dataclass for the Swarm entity. This class is responsible for holding and organizing the individual solutions (i.e. particles), of the optimization problem during the optimization process.

best_n(n: int = 1) → list[Particle | JatParticle][source]

Auxiliary method that returns the best ‘n’ particles with the highest objective function value.

Parameters:

n – the number of the best chromosomes.

Returns:

Return a list with the ‘n’ top particles.

best_particle() → Particle | JatParticle[source]

Auxiliary method that returns the particle with the highest function value. Safeguard with ignoring NaNs.

Returns:

Return the particle with the highest value.

best_positions_as_array() → ndarray[Any, dtype[_ScalarType_co]][source]

Get the particle best positions of all the swarm.

Returns:

A numpy array with all the best positions.

function_values() → ndarray[Any, dtype[_ScalarType_co]][source]

Get the objectives function values of all the swarm.

Returns:

A numpy array (vector) with all the values.

property global_best_index: int

Accessor of the global best index.

Returns:

the index (int) of the best.

property has_categorical: bool

Accessor (getter) of the ‘has_categorical’ flag.

Returns:

true if the data block is CATEGORICAL.

property population: list[Particle | JatParticle]

Accessor of the population list of the swarm.

Returns:

the list (of particles) of the swarm.

position_at(index: int) → _SupportsArray[dtype[Any]] | _NestedSequence[_SupportsArray[dtype[Any]]] | bool | int | float | complex | str | bytes | _NestedSequence[bool | int | float | complex | str | bytes][source]

Get the position vector of an individual particle of the swarm.

Parameters:

index – Position of the individual in the population.

Returns:

The position vector (array).

positions_as_array() → ndarray[Any, dtype[_ScalarType_co]][source]

Get the particle positions of all the swarm.

Returns:

A numpy array with all the positions.

positions_as_list() → list[source]

Get the particle positions of all the swarm.

Returns:

A list with all the positions.

set_positions(new_positions: ndarray[Any, dtype[_ScalarType_co]]) → None[source]

Sets the positions of the particles in the swarm.

Parameters:

new_positions – (NDArray) the new positions of the particles.

Returns:

None.

property size: int

Returns the size (length) of the population. Note that the size of the population is not expected to change during the optimization.

Returns:

(int) the length of the population.

update_local_best() → None[source]

Update the particles in the swarm to their local best values and positions.

Returns:

None.

value_at(index: int) → float[source]

Get the function value of an individual particle of the swarm.

Parameters:

index – Position of the individual in the population.

Returns:

The function value (float).

Module contents