/**
 * Scalar (1-D) function minimization via Brent's method.
 *
 * Distinct from root-finding (`findRoot`, `newton`, `secant`, `halley`, which
 * solve f(x) = 0): this locates a local minimizer of f over a bounded
 * interval. Combines golden-section search (guaranteed, slow) with parabolic
 * interpolation (fast near a smooth minimum) — the classic algorithm of
 * Brent (1973) / Numerical Recipes §10.2, and the same method behind
 * `scipy.optimize.minimize_scalar(method='bounded')`.
 *
 * @packageDocumentation
 */
type f64 = number;
type i32 = number;
/**
 * Options for {@link minimizeScalar}.
 */
export interface MinimizeScalarOptions {
    /**
     * Interval `[a, b]` to search over (bounded Brent). If omitted, defaults to
     * `[-10, 10]` — a generic finite bracket; supply an explicit bracket for
     * functions whose minimum may lie outside that range.
     */
    bracket?: [f64, f64];
    /** Convergence tolerance on the interval width (default 1e-8). */
    tol?: f64;
    /** Maximum iterations (default 100). */
    maxIter?: i32;
}
/**
 * Result of {@link minimizeScalar}.
 */
export interface MinimizeScalarResult {
    /** The minimizer. */
    x: f64;
    /** f(x) at the minimizer. */
    fval: f64;
}
/**
 * Minimize a scalar function `f: R -> R` over a bounded interval using
 * Brent's method (golden-section search + parabolic interpolation).
 *
 * If `opts.bracket` is not supplied, the default search interval is
 * `[-10, 10]`.
 *
 * @param f - Function to minimize
 * @param opts - Options (bracket, tol, maxIter)
 * @returns `{ x, fval }` — the minimizer and its function value
 *
 * @example
 * minimizeScalar(x => (x - 2) ** 2) // => { x: ~2, fval: ~0 }
 * minimizeScalar(Math.sin, { bracket: [0, 2 * Math.PI] }) // => { x: ~3pi/2, fval: ~-1 }
 */
export declare function minimizeScalar(f: (x: f64) => f64, opts?: MinimizeScalarOptions): MinimizeScalarResult;
export {};
//# sourceMappingURL=minimize-scalar.d.ts.map