Crate forge_core 

forge-core

A deterministic metaheuristic optimization substrate in Rust. forge is to optimization what surtgis-core is to raster data: a shared engine that several tools in the ecosystem consume rather than an application in itself.

Every optimizer speaks one interface — the Problem trait — minimizes by convention, counts objective evaluations as its budget, rejects non-finite candidates, and is reproducible for a given seed (the portfolio’s certified-determinism seal).

What sets it apart

Generic Rust optimization crates exist (argmin, optirustic). forge’s value is not to reimplement them but to provide:

1. Geoscientific global optimizers the generic libraries omit — DDS and SCE-UA — central to hydrological calibration.
2. Certified determinism via one seedable RNG (Rng).
3. One unified Problem trait for the whole ecosystem.

The DDS and SCE-UA implementations are migrated from rainflow-core, where they were validated against airGR (GR4J calibration NSE 0.7956 vs 0.7957).

Quick start

use forge_core::{Dds, Optimizer, Termination};
use forge_core::testfn::Sphere;

let problem = Sphere::new(5);
let report = Dds::default().optimize(&problem, &Termination::budget(2000));
assert!(report.best_value() < 1e-2);

Maximization (e.g. NSE/KGE in rainflow) wraps the problem so the minimizing core stays the single convention:

use forge_core::{Optimizer, Sce, Termination};
use forge_core::problem::{func, Maximize};

// Maximize a concave bump peaking at x = 2.
let p = Maximize(func(vec![(-10.0, 10.0)], |x| -(x[0] - 2.0).powi(2)));
let report = Sce::default().optimize(&p, &Termination::budget(3000));
assert!((report.best()[0] - 2.0).abs() < 0.1);
assert!(report.best_value_maximized() > -1e-2);

Robust restarts via independent islands (deterministic with or without the rayon feature):

use forge_core::{ensemble, De, Termination};
use forge_core::testfn::Rastrigin;

let problem = Rastrigin::new(3);
let report = ensemble(&De::default(), &problem, &Termination::budget(5000), 8, 42);
assert!(report.best_value() < 1.0);

Multi-objective optimization with NSGA-II returns a Pareto front:

use forge_core::{NsgaII, Termination};
use forge_core::problem::multi_func;

// Schaffer N.1: minimize x² and (x−2)²; Pareto-optimal x ∈ [0, 2].
let sch = multi_func(vec![(-5.0, 5.0)], 2, |x| vec![x[0] * x[0], (x[0] - 2.0).powi(2)]);
let front = NsgaII::default().optimize(&sch, &Termination::budget(6000));
assert!(!front.is_empty());

Re-exports

pub use algo::ensemble;

pub use algo::Algorithm;

pub use algo::Anneal;

pub use algo::Behavioral;

pub use algo::CmaEs;

pub use algo::Dds;

pub use algo::De;

pub use algo::EpsilonLShade;

pub use algo::Glue;

pub use algo::GlueResult;

pub use algo::GlueSample;

pub use algo::LShade;

pub use algo::LSrtde;

pub use algo::Moead;

pub use algo::NsgaII;

pub use algo::NsgaIII;

pub use algo::Optimizer;

pub use algo::Padds;

pub use algo::PaddsSelection;

pub use algo::ParallelTempering;

pub use algo::Pso;

pub use algo::Restart;

pub use algo::RestartCmaEs;

pub use algo::Sa;

pub use algo::SaResult;

pub use algo::Sce;

pub use algo::Schedule;

pub use algo::SmsEmoa;

pub use algo::TemperingResult;

pub use problem::func;

pub use problem::multi_func;

pub use problem::validate;

pub use problem::validate_multi;

pub use problem::BoundsError;

pub use problem::Maximize;

pub use problem::MultiProblem;

pub use problem::Problem;

pub use rng::Rng;

pub use solution::MultiSolution;

pub use solution::ParetoFront;

pub use solution::Report;

pub use solution::Solution;

pub use solution::StopReason;

pub use termination::Termination;

Modules

algo

Optimization algorithms.

constraint

Nonlinear constraint handling.

indicators

Quality indicators for multi-objective fronts.

problem

The unified optimization problem.

rng

Deterministic RNG (SplitMix64), no external dependencies.

solution

Optimization results shared by every algorithm.

termination

Stopping criteria shared by every optimizer.

testfn

Standard optimization test functions, as Problems.