numerics-ode

Research-grade ordinary differential equation (ODE) solvers implemented in pure Rust with zero external dependencies.

Solvers

• Euler — First-order explicit method. Simple but only O(h) accurate.
• RK4 — Classical fourth-order Runge-Kutta. O(h⁴) global error.
• Adams-Bashforth — Second-order two-step explicit multistep method. O(h²).
• Dormand-Prince — Embedded RK4(5) with adaptive step-size control. O(h⁴/⁵).

System support

All solvers support both scalar ODEs (dy/dx = f(x, y) where y: f64) and systems of ODEs (dy/dx = f(x, &y) where y: Vec<f64>).

Example

use numerics_ode::rk4;

let f = |_x: f64, y: f64| y; // dy/dx = y
let (xs, ys) = rk4::solve(&f, 0.0, 1.0, 1.0, 100);
assert!((ys[100] - (1.0_f64).exp()).abs() < 1e-8);