Struct SimulationBuilder

pub struct SimulationBuilder<T: Individual + Send + Sync> { /* private fields */ }

This is a helper struct in order to build (configure) a valid simulation. See builder pattern: https://en.wikipedia.org/wiki/Builder_pattern

Maybe use phantom types, see https://github.com/willi-kappler/darwin-rs/issues/9

Implementations

impl<T: Individual + Send + Sync> SimulationBuilder<T>

This implementation contains all the helper method to build (configure) a valid simulation.

pub fn new() -> SimulationBuilder<T>

Start with this method, it must always be called as the first one. It creates a default simulation with some dummy (but invalid) values.

pub fn iterations(self, iterations: u32) -> SimulationBuilder<T>

Set the total number of iterations for the simulation and thus sets the simulation type to EndIteration. (Only usefull in combination with EndIteration).

pub fn factor(self, factor: f64) -> SimulationBuilder<T>

Set the improvement factor stop criteria for the simulation and thus sets the simulation type to EndFactor. (Only usefull in combination with EndFactor).

pub fn fitness(self, fitness: f64) -> SimulationBuilder<T>

Set the minimum fitness stop criteria for the simulation and thus sets the simulation type to EndFitness. (Only usefull in combination with EndFactor).

pub fn threads(self, threads: usize) -> SimulationBuilder<T>

Sets the number of threads in order to speed up the simulation.

pub fn add_population(self, population: Population<T>) -> SimulationBuilder<T>

Add a population to the simulation.

pub fn add_multiple_populations( self, multiple_populations: Vec<Population<T>>, ) -> SimulationBuilder<T>

Add multiple populations to the simulation.

pub fn share_fittest(self) -> SimulationBuilder<T>

If this option is enabled (default: off), then the fittest individual of all populations is shared between all populations.

pub fn num_of_global_fittest( self, num_of_global_fittest: usize, ) -> SimulationBuilder<T>

How many global fittest should be kept ? (The size of the “high score list”)

pub fn output_every(self, output_every: u32) -> SimulationBuilder<T>

Do not output every time a new individual is found, only every nth time. n == output_every

pub fn share_every(self, share_every: u32) -> SimulationBuilder<T>

If share fittest is enabled and the number share_every of iteration has passed then the fittest individual is shared between all populations

pub fn finalize(self) -> Result<Simulation<T>>

This checks the configuration of the simulation and returns an error or Ok if no errors where found.