Crate cmaes

A standard CMA-ES implementation with multithreaded support. Implemented based on example code here.

The algorithm minimizes the fitness function, so a lower fitness represents a better individual.

Examples

use cmaes::*;

#[derive(Clone)]
struct FitnessDummy;

impl FitnessFunction for FitnessDummy {
    fn get_fitness(&self, parameters: &[f64]) -> f64 {
        // Calculate fitness here

        0.0
    }
}

// See the documentation for CMAESOptions for a list of all options
let options = CMAESOptions::default(2);

// solution will be tuple with a set of optimized parameters, with as low a fitness as
// possible, and its fitness value
let solution = cmaes_loop(&FitnessDummy, options);

Reexports

pub use self::cmaes_loop::cmaes_loop;

pub use self::fitness::FitnessFunction;

pub use self::options::CMAESOptions;

Modules

cmaes_loop
fitness
options	Option types for the CMA-ES algorithm