iterative_solvers/
solver.rs

1//! A trait for iterative solvers.
2
3use crate::{IterSolverError, IterSolverResult};
4use nalgebra::{DMatrix, DVector};
5
6/// A linear system `Ax = b` with a matrix `A` and a right-hand side vector `b`.
7pub struct LinearSystem {
8    mat: DMatrix<f64>,
9    rhs: DVector<f64>,
10}
11
12impl LinearSystem {
13    /// Create a new `LinearSystem` with a matrix `A` and a right-hand side vector `b`.
14    ///
15    /// # Arguments
16    ///
17    /// * `mat` - The matrix `A`.
18    /// * `rhs` - The right-hand side vector `b`.
19    ///
20    /// # Errors
21    ///
22    /// Returns an error if the matrix is not square or if the matrix and the right-hand side vector do not match.
23    ///
24    /// # Examples
25    ///
26    /// ```rust
27    /// use nalgebra::{DMatrix, DVector};
28    /// use iterative_solvers::LinearProblem;
29    ///
30    /// let mat = DMatrix::from_row_slice(2, 2, &[4.0, 1.0, 1.0, 3.0]);
31    /// let rhs = DVector::from_vec(vec![1.0, 2.0]);
32    /// let problem = LinearProblem::new(mat, rhs).unwrap();
33    /// ```
34    pub fn new(mat: DMatrix<f64>, rhs: DVector<f64>) -> IterSolverResult<Self> {
35        if !mat.is_square() {
36            return Err(IterSolverError::DimensionError(format!(
37                "The matrix is not square, whose shape is ({}, {})",
38                mat.shape().0,
39                mat.shape().1
40            )));
41        }
42        if mat.nrows() != rhs.len() {
43            return Err(IterSolverError::DimensionError(format!(
44                "The matrix with order {}, and the rhs with length {}, do not match",
45                mat.nrows(),
46                rhs.len()
47            )));
48        }
49        Ok(Self { mat, rhs })
50    }
51
52    /// Solve the linear problem `Ax = b` using the given solver.
53    ///
54    /// # Arguments
55    ///
56    /// * `solver` - The solver to use.
57    ///
58    /// # Errors
59    ///
60    /// Returns an error if the solver fails to solve the linear problem.
61    ///
62    /// # Examples
63    ///
64    /// ```rust
65    /// use nalgebra::{DMatrix, DVector};
66    /// use iterative_solvers::{LinearProblem, CG};
67    ///
68    /// let mat = DMatrix::from_row_slice(2, 2, &[4.0, 1.0, 1.0, 3.0]);
69    /// let rhs = DVector::from_vec(vec![1.0, 2.0]);
70    /// let problem = LinearProblem::new(mat, rhs).unwrap();
71    /// let solver = CG::new(&mat, &rhs, 1e-6).unwrap();
72    /// let solution = problem.solve(solver).unwrap();
73    /// ```
74    pub fn solve(&self, solver: impl IterativeSolver) -> IterSolverResult<DVector<f64>> {
75        solver.solve(&self.mat, &self.rhs)
76    }
77}
78
79/// A trait for iterative solvers.
80pub trait IterativeSolver {
81    /// Solve the linear problem `Ax = b` using the given solver.
82    ///
83    /// # Arguments
84    ///
85    /// * `mat` - The matrix `A`.
86    /// * `rhs` - The right-hand side vector `b`.
87    ///
88    /// # Returns
89    ///
90    /// A `DVector` containing the solution to the linear problem.
91    fn solve(&self, mat: &DMatrix<f64>, rhs: &DVector<f64>) -> IterSolverResult<DVector<f64>>;
92}
iterative_solvers/solver.rs

iterative_solvers/
solver.rs
