Module ojalgo
Package org.ojalgo.matrix.task.iterative

Class GaussSeidelSolver

  • All Implemented Interfaces:
    MatrixTask<java.lang.Double>, SolverTask<java.lang.Double>
    public final class GaussSeidelSolver
extends IterativeSolverTask
    Stationary Gauss–Seidel iteration for solving [A][x]=[b] with non-zero diagonal entries.

    Convergence

    • Converges for strictly diagonally dominant systems and for symmetric positive-definite (SPD) matrices.
    • Behaviour depends on ordering and scaling; preconditioning is not applied in this stationary method.
    Configuration
    • Ignores any configured Preconditioner; use the relaxation factor to control convergence speed.

    When to use

    • As a simple in-place fixed-point iteration when sequential updates are acceptable.
    • Prefer over fully synchronous updates when in-place coupling improves convergence.
    • For large SPD problems needing faster convergence, Krylov methods often perform better.
    • If fully synchronous updates or trivial parallelism are required, consider a synchronous fixed-point method.