srfm::portfolio::GeodesicSolver Class Reference

#include <geodesic_path.hpp>

Static Public Member Functions
static Geodesic	solve (const PortfolioState &start, const PortfolioState &end, int n_steps, double lambda)

Detailed Description

Computes the geodesic between two portfolio states under a concentration penalty Lagrangian.

The analytical solution: omega = sqrt(2 * lambda) A_i = start.weights[i] B_i = (end.weights[i] - A_i * cos(omega)) / sin(omega) if sin(omega) != 0 (linear interpolation fallback when omega is near zero) w_i(t_norm) = A_i * cos(omega * t_norm) + B_i * sin(omega * t_norm)

where t_norm in [0, 1] is the normalised time along the geodesic.

Definition at line 99 of file geodesic_path.hpp.

Member Function Documentation

solve()

Geodesic srfm::portfolio::GeodesicSolver::solve ( const PortfolioState &  start, const PortfolioState &  end, int  n_steps, double  lambda  )

static

Solve for the geodesic between start and end.

Parameters

start	Initial portfolio state.
end	Target portfolio state. Must have the same dimension as start.
n_steps	Number of interior discretisation steps (>= 1). Total waypoints = n_steps + 1 (endpoints included).
lambda	Concentration penalty coefficient (>= 0). lambda = 0 → straight-line (flat spacetime) geodesic.

Returns

Geodesic with (n_steps + 1) waypoints.

Exceptions

std::invalid_argument

if dimensions mismatch or n_steps < 1.

Definition at line 77 of file geodesic_path.cpp.

---

The documentation for this class was generated from the following files:

• include/srfm/geodesic_path.hpp
• src/geodesic_path.cpp