Struct NeumannPreconditioner 

pub struct NeumannPreconditioner { /* private fields */ }

Neumann series preconditioner.

Approximates A^{-1} by the truncated Neumann series: M^{-1} = alpha * sum_{k=0}^{d} (I - alpha * A)^k

Convergent when ||I - alpha * A|| < 1. The scaling alpha is chosen automatically as 1 / ||A||_inf (infinity norm).

Implementations

impl NeumannPreconditioner

pub fn from_csr( csr: &CsrMatrix<f64>, degree: usize, alpha: Option<f64>, ) -> SparseResult<Self>

Build a Neumann series preconditioner from a CSR matrix.

Arguments

• csr - The sparse matrix in CSR format (should be well-conditioned).
• degree - Number of terms in the Neumann series (1-10 typical).
• alpha - Scaling factor. Pass None for automatic estimation.

pub fn apply(&self, r: &[f64]) -> SparseResult<Vec<f64>>

Apply the Neumann preconditioner: z = M^{-1} * r.

Evaluates z = alpha * sum_{k=0}^{d} (I - alpha*A)^k * r using Horner’s scheme in degree SpMV operations.

pub fn alpha(&self) -> f64

Return the scaling factor alpha.

pub fn degree(&self) -> usize

Return the polynomial degree.

pub fn size(&self) -> usize

Return the matrix dimension.