Struct MeanVarianceOptimizer 

pub struct MeanVarianceOptimizer;

Mean-variance portfolio optimizer using gradient descent.

Implementations

impl MeanVarianceOptimizer

pub fn portfolio_return(weights: &[f64], returns: &[f64]) -> f64

Compute expected portfolio return: dot product of weights and returns.

pub fn portfolio_variance(weights: &[f64], cov: &[Vec<f64>]) -> f64

Compute portfolio variance: w^T Σ w.

pub fn portfolio_volatility(weights: &[f64], cov: &[Vec<f64>]) -> f64

Compute portfolio volatility: sqrt(w^T Σ w).

pub fn gradient_step( weights: &mut Vec<f64>, returns: &[f64], cov: &[Vec<f64>], target_return: f64, lr: f64, )

Perform one gradient step minimizing variance subject to a target return.

Uses the gradient of Var(w) - lambda * (Return(w) - target) and projects onto the unit simplex afterwards.

pub fn apply_constraints( weights: &mut Vec<f64>, constraints: &[OptimizationConstraint], )

Apply constraints to a weight vector (projection onto constraint set).

pub fn maximize_sharpe( returns: &[f64], cov: &[Vec<f64>], rf_rate: f64, constraints: &[OptimizationConstraint], ) -> OptimizationResult

Maximize Sharpe ratio via gradient ascent.

Iterates gradient ascent on the Sharpe ratio objective: (Return(w) - rf) / Volatility(w).

pub fn minimize_variance( returns: &[f64], cov: &[Vec<f64>], target_return: f64, constraints: &[OptimizationConstraint], ) -> OptimizationResult

Minimize portfolio variance subject to a target return.

pub fn efficient_frontier( returns: &[f64], cov: &[Vec<f64>], n_points: usize, rf_rate: f64, ) -> Vec<EfficientFrontierPoint>

Generate the efficient frontier by sweeping target returns.

Produces n_points EfficientFrontierPoint values spanning the range from the minimum expected return to the maximum expected return.