Struct De 

pub struct De { /* private fields */ }

Stateful Differential Evolution optimizer.

Use optimize for an in-process objective or ask/tell when another system evaluates a population.

Implementations

impl De

pub fn new( fitfun: Fitness, guess: &[f64], sigma: &[f64], ints: Option<Vec<bool>>, p: &DeParams, ) -> Self

Build a DE optimizer. guess/sigma empty ⇒ uniform sampling in the box; non-empty ⇒ normal sampling around guess. ints marks discrete coordinates (length dim) or is None.

pub fn dim(&self) -> usize

Number of decision variables.

pub fn popsize(&self) -> usize

Number of individuals evaluated per generation.

pub fn stop(&self) -> i32

Current termination code, or zero while the optimizer can continue.

pub fn optimize(&mut self, obj: &impl Objective) -> DeResult

Run the serial generational loop against an in-process objective.

pub fn ask(&mut self) -> Vec<Vec<f64>>

Ask for a full population of candidate rows.

pub fn tell(&mut self, ys: &[f64]) -> i32

Tell fitness values for the population returned by ask.

pub fn population(&self) -> Vec<Vec<f64>>

Return a decoded snapshot of the current population.

pub fn result(&self) -> DeResult

Return the current best result for an ask/tell run.