scivex-optim 0.1.0

//! Quadratic programming via the active set method.
//!
//! Solves problems of the form:
//!
//! ```text
//! minimize    0.5 x^T H x + c^T x
//! subject to  A_ub x <= b_ub
//! ```
//!
//! where `H` is symmetric positive semi-definite.

pub mod active_set;

pub use active_set::quadprog;

use scivex_core::Float;

/// Result of a quadratic programming solve.
///
/// # Examples
///
/// ```
/// # use scivex_optim::quadprog::quadprog;
/// // minimize 0.5*(x^2+y^2)  s.t. x+y <= 1
/// let result = quadprog(
///     &[vec![1.0_f64, 0.0], vec![0.0, 1.0]],
///     &[0.0, 0.0],
///     &[vec![1.0, 1.0]], &[1.0], 100,
/// ).unwrap();
/// assert!(result.converged);
/// ```
#[cfg_attr(
    feature = "serde-support",
    derive(serde::Serialize, serde::Deserialize)
)]
#[derive(Debug, Clone)]
pub struct QpResult<T: Float> {
    /// Optimal decision variables.
    pub x: Vec<T>,
    /// Optimal objective value (`0.5 x^T H x + c^T x`).
    pub fun: T,
    /// Number of active set iterations.
    pub iterations: usize,
    /// Whether the solver converged to an optimal solution.
    pub converged: bool,
}