Iterative_process

Iterative_process

Iterative_process(x_0, tol=1e-05, lambda_min=0.1, lambda_max=1.0, n_ini_relax=2, rel_vel=0.9)

Iterative process using the fixed-point method with adaptive relaxation coefficient

Used by the components to converge a value coupled with other components (e.g. the temperature or humidity of a space) within the iterations of a time step:

it = Iterative_process(x_0, tol=0.01)
# in each iteration, with the value proposed by the component's equations
x = it.estimate_next_x(proposed_x)
converged = it.converged()

Create new iterative process

Parameters:

Name Type Description Default
x_0 float

Initial value

required
tol float

convergence tolerance of the residual. Defaults to 1e-5.

1e-05
lambda_min float

minimum relaxation coefficient. Defaults to 0.1.

0.1
lambda_max float

maximum (and initial) relaxation coefficient. Defaults to 1.0.

1.0
n_ini_relax int

Number of iterations in which adaptive relaxation is initiated, must be greater than 1. Defaults to 2.

2
rel_vel float

factor that reduces (or, divided, recovers) the relaxation coefficient. Defaults to 0.9.

0.9

converged

converged()

The residual of the last iteration is below the tolerance

Returns:

Type Description
bool

converged

estimate_next_x

estimate_next_x(proposed_x)

Relaxed next value from the value proposed by the fixed-point equation

Parameters:

Name Type Description Default
proposed_x float

value given by the equations from the last value

required

Returns:

Type Description
float

next value, (1 - lambda) * x + lambda * proposed_x

set_next_x

set_next_x(next_x)

Set the next value directly, without relaxation

Parameters:

Name Type Description Default
next_x float

next value

required