Struct SolverConfig 

pub struct SolverConfig {
    pub prev_value_variation: f64,
    pub epsilon: f64,
    pub max_ulps: u32,
    pub maxiter: usize,
    pub finite_difference_stepwidth: f64,
    pub dampening: f64,
}

Solver parameters used in the NetworkAnalysis::solve function.

The default values given for each field are used in the implementation of the Default trait.

Fields

prev_value_variation: f64

This value is used to create a “zeroth” iteration value which is necessary to start the first iteration. This is a heuristic value which can be varied if the solver does not converge. Default is 0.9.

epsilon: f64

The Newton-Raphson algorithmus used in the implementations of NetworkAnalysis uses the ulps_eq from the approx crate in order to determine whether the root has been found with sufficient precision (i.e., f(x_0) == 0 is approximately fulfilled). ulps_eq uses both an absolute difference check (using the epsilon field) and an ULPs (using max_ulps field) comparison. For the former check, is used. Default is the square root of the machine precision (std::f64::EPSILON.sqrt()).

max_ulps: u32

See epsilon. Default is 4.

maxiter: usize

Maximum number of iterations of the underlying Newton-Raphson solver. When maxiter is exceeded, the solver terminates with an error. Default is 200.

finite_difference_stepwidth: f64

If no custom Jacobian function is given, the Jacobian is calculated as a finite difference approximation (see [1]) This field finite_difference_stepwidth corresponds to the stepwidth h. Default is the square root of the machine precision (std::f64::EPSILON.sqrt()).

1. Pauletti, Ruy & Almeida Neto, Edgard: A finite-difference approximation to Newton´s Method Jacobian Matrices. IABSE-IASS Symposium London 2011).

dampening: f64

As described in [2], the dampening factor tau can be used to increase the convergence probability at the cost of higher iteration time. Default is 1.0 (no dampening).

2. Okawa, Hirotada et al.: The W4 method: A new multi-dimensional root-finding scheme for nonlinear systems of equations. Applied Numerical Mathematics 183 (2023) 157-172.

Trait Implementations

impl Clone for SolverConfig

fn clone(&self) -> SolverConfig

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for SolverConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for SolverConfig

fn default() -> Self

See the docstring of SolverConfig for the default parameters values.