Integrator Configuration¶

Configuration classes and types for controlling integrator behavior.

IntegratorConfig ¶

IntegratorConfig(abs_tol: float = 1e-06, rel_tol: float = 0.001, initial_step: float = None, min_step: float = Ellipsis, max_step: float = Ellipsis, step_safety_factor: float = Ellipsis, min_step_scale_factor: float = Ellipsis, max_step_scale_factor: float = Ellipsis, max_step_attempts: int = 10, fixed_step_size: float = None)

Configuration for numerical integrators.

Controls error tolerances, step size limits, and other integration parameters.

Parameters:

Name	Type	Description	Default
`abs_tol`	`float`	Absolute error tolerance. Defaults to 1e-6.	`1e-06`
`rel_tol`	`float`	Relative error tolerance. Defaults to 1e-3.	`0.001`
`initial_step`	`float`	Initial step size. Defaults to None (auto).	`None`
`min_step`	`float`	Minimum step size. Defaults to 1e-12.	`Ellipsis`
`max_step`	`float`	Maximum step size. Defaults to 900.0.	`Ellipsis`
`step_safety_factor`	`float`	Safety factor for step control. Defaults to 0.9.	`Ellipsis`
`min_step_scale_factor`	`float`	Minimum step scaling. Defaults to 0.2.	`Ellipsis`
`max_step_scale_factor`	`float`	Maximum step scaling. Defaults to 10.0.	`Ellipsis`
`max_step_attempts`	`int`	Maximum step attempts. Defaults to 10.	`10`
`fixed_step_size`	`float`	Fixed step size for fixed-step integrators. Defaults to None.	`None`

Example

import brahe as bh

# Create default configuration
config = bh.IntegratorConfig()

# Create fixed-step configuration
config = bh.IntegratorConfig.fixed_step(1.0)

# Create adaptive configuration with custom tolerances
config = bh.IntegratorConfig.adaptive(abs_tol=1e-9, rel_tol=1e-6)

# Customize configuration
config = bh.IntegratorConfig(
    abs_tol=1e-8,
    rel_tol=1e-6,
    max_step=100.0,
    min_step=0.001
)

Initialize instance.

abs_tol `property` ¶

abs_tol: Any

Get absolute error tolerance.

initial_step `property` ¶

initial_step: Any

Get initial step size.

max_step `property` ¶

max_step: Any

Get maximum step size.

max_step_attempts `property` ¶

max_step_attempts: Any

Get maximum step attempts.

max_step_scale_factor `property` ¶

max_step_scale_factor: Any

Get maximum step scale factor.

min_step `property` ¶

min_step: Any

Get minimum step size.

min_step_scale_factor `property` ¶

min_step_scale_factor: Any

Get minimum step scale factor.

rel_tol `property` ¶

rel_tol: Any

Get relative error tolerance.

step_safety_factor `property` ¶

step_safety_factor: Any

Get step safety factor.

adaptive `builtin` ¶

adaptive(abs_tol: float, rel_tol: float) -> IntegratorConfig

Create a configuration for adaptive-step integration.

Parameters:

Name	Type	Description	Default
`abs_tol`	`float`	Absolute error tolerance	required
`rel_tol`	`float`	Relative error tolerance	required

Returns:

Name	Type	Description
`IntegratorConfig`	`IntegratorConfig`	Configuration for adaptive-step integration

Example

import brahe as bh
config = bh.IntegratorConfig.adaptive(1e-9, 1e-6)

fixed_step `builtin` ¶

fixed_step(step_size: float) -> IntegratorConfig

Create a configuration for fixed-step integration.

Parameters:

Name	Type	Description	Default
`step_size`	`float`	Fixed timestep in seconds	required

Returns:

Name	Type	Description
`IntegratorConfig`	`IntegratorConfig`	Configuration for fixed-step integration

Example

import brahe as bh
config = bh.IntegratorConfig.fixed_step(1.0)

AdaptiveStepResult ¶

AdaptiveStepResult()

Result from an adaptive integration step.

Contains the new state, actual timestep used, error estimate, and suggested next step.

Example

import brahe as bh
import numpy as np

def dynamics(t, state):
    return np.array([state[1], -state[0]])

integrator = bh.RKF45Integrator(2, dynamics)
state = np.array([1.0, 0.0])

result = integrator.step(0.0, state, 0.1)
print(f"New state: {result.state}")
print(f"Step used: {result.dt_used}")
print(f"Error: {result.error_estimate}")
print(f"Next step: {result.dt_next}")

Initialize instance.

dt_next `property` ¶

dt_next: float

Get the suggested next timestep.

Returns:

Name	Type	Description
`float`	`float`	Suggested timestep for next iteration

dt_used `property` ¶

dt_used: float

Get the actual timestep used.

Returns:

Name	Type	Description
`float`	`float`	Timestep actually used (may be smaller than requested)

error_estimate `property` ¶

error_estimate: float

Get the estimated truncation error.

Returns:

Type	Description
`float`	float \| None: Estimated local truncation error (None for fixed-step integrators)

state `property` ¶

state: ndarray

Get the new state vector.

Returns:

Name	Type	Description
`ndarray`	`ndarray`	State vector at time t + dt_used

Integrator Configuration¶