Trait metaheuristics_nature::ObjFunc

pub trait ObjFunc {
    type Result;
    fn fitness<'a, A>(&self, gen: u32, v: A) -> f64
    where
        A: AsArray<'a, f64>;
    fn result<'a, A>(&self, v: A) -> Self::Result
    where
        A: AsArray<'a, f64>;
    fn ub(&self) -> ArrayView1<'_, f64>;
    fn lb(&self) -> ArrayView1<'_, f64>;
}

The base of the objective function.

For example:

use metaheuristics_nature::ObjFunc;
use ndarray::{AsArray, ArrayView1, Array1};
struct MyFunc(Array1<f64>, Array1<f64>);
impl MyFunc {
    fn new() -> Self { Self(Array1::zeros(3), Array1::ones(3) * 50.) }
}
impl ObjFunc for MyFunc {
    type Result = f64;
    fn fitness<'a, A>(&self, gen: u32, v: A) -> f64
    where
        A: AsArray<'a, f64>,
    {
        let v = v.into();
        v[0] * v[0] + v[1] * v[1] + v[2] * v[2]
    }
    fn result<'a, A>(&self, v: A) -> Self::Result
    where
        A: AsArray<'a, f64>
    {
        self.fitness(0, v)
    }
    fn ub(&self) -> ArrayView1<f64> { self.1.view() }
    fn lb(&self) -> ArrayView1<f64> { self.0.view() }
}

The objective function returns fitness value that used to evaluate the objective.

The lower bound and upper bound represents the number of variables at the same time.

This trait is designed as immutable.

Associated Types

type Result

The result type.

Required methods

fn fitness<'a, A>(&self, gen: u32, v: A) -> f64 where
    A: AsArray<'a, f64>, 

Return fitness, the smaller value represents good.

fn result<'a, A>(&self, v: A) -> Self::Result where
    A: AsArray<'a, f64>, 

Return the final result of the problem.

fn ub(&self) -> ArrayView1<'_, f64>

Get upper bound.

fn lb(&self) -> ArrayView1<'_, f64>

Get lower bound.

Implementors