LinearSolverPolicy

class probnum.linalg.solvers.policies.LinearSolverPolicy

Bases: ABC

Policy of a (probabilistic) linear solver.

The policy \(\pi(s \mid \mathsf{A}, \mathsf{H}, \mathsf{x}, A, b)\) of a linear solver returns a vector to probe the linear system with, typically via multiplication, resulting in an observation. Policies can either be deterministic or stochastic depending on the application.

See also

ConjugateDirectionsPolicy

Policy returning \(A\)-conjugate actions.

RandomUnitVectorPolicy

Policy returning random standard unit vectors.

Methods Summary

__call__(solver_state[, rng])

Return an action for a given solver state.

Methods Documentation

abstract __call__(solver_state, rng=None)

Return an action for a given solver state.

Parameters:

• solver_state (LinearSolverState) – Current state of the linear solver.

• rng (Generator | None) – Random number generator.

Returns:

Next action to take.

Return type:

action