Coverage for pySDC/implementations/convergence_controller_classes/adaptivity.py: 95%
257 statements
« prev ^ index » next coverage.py v7.5.0, created at 2024-04-29 09:02 +0000
1import numpy as np
2from pySDC.core.ConvergenceController import ConvergenceController, Status
3from pySDC.implementations.convergence_controller_classes.step_size_limiter import (
4 StepSizeLimiter,
5)
8class AdaptivityBase(ConvergenceController):
9 """
10 Abstract base class for convergence controllers that implement adaptivity based on arbitrary local error estimates
11 and update rules.
12 """
14 def setup(self, controller, params, description, **kwargs):
15 """
16 Define default parameters here.
18 Default parameters are:
19 - control_order (int): The order relative to other convergence controllers
20 - beta (float): The safety factor
22 Args:
23 controller (pySDC.Controller): The controller
24 params (dict): The params passed for this specific convergence controller
25 description (dict): The description object used to instantiate the controller
27 Returns:
28 (dict): The updated params dictionary
29 """
30 defaults = {
31 "control_order": -50,
32 "beta": 0.9,
33 }
34 from pySDC.implementations.hooks.log_step_size import LogStepSize
36 controller.add_hook(LogStepSize)
38 from pySDC.implementations.convergence_controller_classes.check_convergence import CheckConvergence
40 self.communicate_convergence = CheckConvergence.communicate_convergence
42 return {**defaults, **super().setup(controller, params, description, **kwargs)}
44 def dependencies(self, controller, description, **kwargs):
45 """
46 Load step size limiters here, if they are desired.
48 Args:
49 controller (pySDC.Controller): The controller
50 description (dict): The description object used to instantiate the controller
52 Returns:
53 None
54 """
55 step_limiter_keys = ['dt_min', 'dt_max', 'dt_slope_min', 'dt_slope_max']
56 available_keys = [me for me in step_limiter_keys if me in self.params.__dict__.keys()]
58 if len(available_keys) > 0:
59 step_limiter_params = {key: self.params.__dict__[key] for key in available_keys}
60 controller.add_convergence_controller(StepSizeLimiter, params=step_limiter_params, description=description)
62 if self.params.useMPI:
63 self.prepare_MPI_logical_operations()
65 return None
67 def get_new_step_size(self, controller, S, **kwargs):
68 """
69 Determine a step size for the next step from an estimate of the local error of the current step.
71 Args:
72 controller (pySDC.Controller): The controller
73 S (pySDC.Step): The current step
75 Returns:
76 None
77 """
78 raise NotImplementedError("Please implement a rule for updating the step size!")
80 def compute_optimal_step_size(self, beta, dt, e_tol, e_est, order):
81 """
82 Compute the optimal step size for the current step based on the order of the scheme.
83 This function can be called from `get_new_step_size` for various implementations of adaptivity, but notably not
84 all! We require to know the order of the error estimate and if we do adaptivity based on the residual, for
85 instance, we do not know that and we can't use this function.
87 Args:
88 beta (float): Safety factor
89 dt (float): Current step size
90 e_tol (float): The desired tolerance
91 e_est (float): The estimated local error
92 order (int): The order of the local error estimate
94 Returns:
95 float: The optimal step size
96 """
97 return beta * dt * (e_tol / e_est) ** (1.0 / order)
99 def get_local_error_estimate(self, controller, S, **kwargs):
100 """
101 Get the local error estimate for updating the step size.
102 It does not have to be an error estimate, but could be the residual or something else.
104 Args:
105 controller (pySDC.Controller): The controller
106 S (pySDC.Step): The current step
108 Returns:
109 float: The error estimate
110 """
111 raise NotImplementedError("Please implement a way to get the local error")
113 def determine_restart(self, controller, S, **kwargs):
114 """
115 Check if the step wants to be restarted by comparing the estimate of the local error to a preset tolerance
117 Args:
118 controller (pySDC.Controller): The controller
119 S (pySDC.Step): The current step
121 Returns:
122 None
123 """
124 if S.status.iter >= S.params.maxiter:
125 e_est = self.get_local_error_estimate(controller, S)
126 if e_est >= self.params.e_tol:
127 # see if we try to avoid restarts
128 if self.params.get('avoid_restarts'):
129 more_iter_needed = max([L.status.iter_to_convergence for L in S.levels])
130 k_final = S.status.iter + more_iter_needed
131 rho = max([L.status.contraction_factor for L in S.levels])
132 coll_order = S.levels[0].sweep.coll.order
134 if rho > 1:
135 S.status.restart = True
136 self.log(f"Convergence factor = {rho:.2e} > 1 -> restarting", S)
137 elif k_final > 2 * S.params.maxiter:
138 S.status.restart = True
139 self.log(
140 f"{more_iter_needed} more iterations needed for convergence -> restart is more efficient", S
141 )
142 elif k_final > coll_order:
143 S.status.restart = True
144 self.log(
145 f"{more_iter_needed} more iterations needed for convergence -> restart because collocation problem would be over resolved",
146 S,
147 )
148 else:
149 S.status.force_continue = True
150 self.log(f"{more_iter_needed} more iterations needed for convergence -> no restart", S)
151 else:
152 S.status.restart = True
153 self.log(f"Restarting: e={e_est:.2e} >= e_tol={self.params.e_tol:.2e}", S)
155 return None
158class AdaptivityForConvergedCollocationProblems(AdaptivityBase):
159 def dependencies(self, controller, description, **kwargs):
160 """
161 Load interpolation between restarts.
163 Args:
164 controller (pySDC.Controller): The controller
165 description (dict): The description object used to instantiate the controller
167 Returns:
168 None
169 """
170 super().dependencies(controller, description, **kwargs)
172 if self.params.interpolate_between_restarts:
173 from pySDC.implementations.convergence_controller_classes.interpolate_between_restarts import (
174 InterpolateBetweenRestarts,
175 )
177 controller.add_convergence_controller(
178 InterpolateBetweenRestarts,
179 description=description,
180 params={},
181 )
182 if self.params.interpolate_between_restarts:
183 self.interpolator = controller.convergence_controllers[-1]
184 return None
186 def get_convergence(self, controller, S, **kwargs):
187 raise NotImplementedError("Please implement a way to check if the collocation problem is converged!")
189 def setup(self, controller, params, description, **kwargs):
190 """
191 Add a default value for control order to the parameters.
193 Args:
194 controller (pySDC.Controller): The controller
195 params (dict): Parameters for the convergence controller
196 description (dict): The description object used to instantiate the controller
198 Returns:
199 dict: Updated parameters
200 """
201 defaults = {
202 'restol_rel': None,
203 'e_tol_rel': None,
204 'restart_at_maxiter': True,
205 'restol_min': 1e-12,
206 'restol_max': 1e-5,
207 'factor_if_not_converged': 4.0,
208 'residual_max_tol': 1e9,
209 'maxiter': description['sweeper_params'].get('maxiter', 99),
210 'interpolate_between_restarts': True,
211 'abort_at_growing_residual': True,
212 **super().setup(controller, params, description, **kwargs),
213 }
214 if defaults['restol_rel']:
215 description['level_params']['restol'] = min(
216 [max([defaults['restol_rel'] * defaults['e_tol'], defaults['restol_min']]), defaults['restol_max']]
217 )
218 elif defaults['e_tol_rel']:
219 description['level_params']['e_tol'] = min([max([defaults['e_tol_rel'] * defaults['e_tol'], 1e-10]), 1e-5])
221 if defaults['restart_at_maxiter']:
222 defaults['maxiter'] = description['step_params'].get('maxiter', 99)
224 self.res_last_iter = np.inf
226 return defaults
228 def determine_restart(self, controller, S, **kwargs):
229 if self.get_convergence(controller, S, **kwargs):
230 self.res_last_iter = np.inf
232 if self.params.restart_at_maxiter and S.levels[0].status.residual > S.levels[0].params.restol:
233 self.trigger_restart_upon_nonconvergence(S)
234 elif self.get_local_error_estimate(controller, S, **kwargs) > self.params.e_tol:
235 S.status.restart = True
236 elif (
237 S.status.time_size == 1
238 and self.res_last_iter < S.levels[0].status.residual
239 and S.status.iter > 0
240 and self.params.abort_at_growing_residual
241 ):
242 self.trigger_restart_upon_nonconvergence(S)
243 elif S.levels[0].status.residual > self.params.residual_max_tol:
244 self.trigger_restart_upon_nonconvergence(S)
246 if self.params.useMPI:
247 self.communicate_convergence(self, controller, S, **kwargs)
249 self.res_last_iter = S.levels[0].status.residual * 1.0
251 def trigger_restart_upon_nonconvergence(self, S):
252 S.status.restart = True
253 S.status.force_done = True
254 for L in S.levels:
255 L.status.dt_new = L.params.dt / self.params.factor_if_not_converged
256 self.log(
257 f'Collocation problem not converged. Reducing step size to {L.status.dt_new:.2e}',
258 S,
259 )
260 if self.params.interpolate_between_restarts:
261 self.interpolator.status.skip_interpolation = True
264class Adaptivity(AdaptivityBase):
265 """
266 Class to compute time step size adaptively based on embedded error estimate.
268 We have a version working in non-MPI pipelined SDC, but Adaptivity requires you to know the order of the scheme,
269 which you can also know for block-Jacobi, but it works differently and it is only implemented for block
270 Gauss-Seidel so far.
272 There is an option to reduce restarts if continued iterations could yield convergence in fewer iterations than
273 restarting based on an estimate of the contraction factor.
274 Since often only one or two more iterations suffice, this can boost efficiency of adaptivity significantly.
275 Notice that the computed step size is not effected.
276 Be aware that this does not work when Hot Rod is enabled, since that requires us to know the order of the scheme in
277 more detail. Since we reset to the second to last sweep before moving on, we cannot continue to iterate.
278 Set the reduced restart up by setting a boolean value for "avoid_restarts" in the parameters for the convergence
279 controller.
280 The behaviour in multi-step SDC is not well studied and it is unclear if anything useful happens there.
281 """
283 def setup(self, controller, params, description, **kwargs):
284 """
285 Define default parameters here.
287 Default parameters are:
288 - control_order (int): The order relative to other convergence controllers
289 - beta (float): The safety factor
291 Args:
292 controller (pySDC.Controller): The controller
293 params (dict): The params passed for this specific convergence controller
294 description (dict): The description object used to instantiate the controller
296 Returns:
297 (dict): The updated params dictionary
298 """
299 defaults = {
300 "embedded_error_flavor": 'standard',
301 }
302 return {**defaults, **super().setup(controller, params, description, **kwargs)}
304 def dependencies(self, controller, description, **kwargs):
305 """
306 Load the embedded error estimator.
308 Args:
309 controller (pySDC.Controller): The controller
310 description (dict): The description object used to instantiate the controller
312 Returns:
313 None
314 """
315 from pySDC.implementations.convergence_controller_classes.estimate_embedded_error import EstimateEmbeddedError
317 super().dependencies(controller, description, **kwargs)
319 controller.add_convergence_controller(
320 EstimateEmbeddedError.get_implementation(self.params.embedded_error_flavor, self.params.useMPI),
321 description=description,
322 )
324 # load contraction factor estimator if necessary
325 if self.params.get('avoid_restarts'):
326 from pySDC.implementations.convergence_controller_classes.estimate_contraction_factor import (
327 EstimateContractionFactor,
328 )
330 params = {'e_tol': self.params.e_tol}
331 controller.add_convergence_controller(EstimateContractionFactor, description=description, params=params)
332 return None
334 def check_parameters(self, controller, params, description, **kwargs):
335 """
336 Check whether parameters are compatible with whatever assumptions went into the step size functions etc.
337 For adaptivity, we need to know the order of the scheme.
339 Args:
340 controller (pySDC.Controller): The controller
341 params (dict): The params passed for this specific convergence controller
342 description (dict): The description object used to instantiate the controller
344 Returns:
345 bool: Whether the parameters are compatible
346 str: The error message
347 """
348 if description["level_params"].get("restol", -1.0) >= 0:
349 return (
350 False,
351 "Adaptivity needs constant order in time and hence restol in the step parameters has to be \
352smaller than 0!",
353 )
355 if controller.params.mssdc_jac:
356 return (
357 False,
358 "Adaptivity needs the same order on all steps, please activate Gauss-Seidel multistep mode!",
359 )
361 if "e_tol" not in params.keys():
362 return (
363 False,
364 "Adaptivity needs a local tolerance! Please pass `e_tol` to the parameters for this convergence controller!",
365 )
367 return True, ""
369 def get_new_step_size(self, controller, S, **kwargs):
370 """
371 Determine a step size for the next step from an embedded estimate of the local error of the current step.
373 Args:
374 controller (pySDC.Controller): The controller
375 S (pySDC.Step): The current step
377 Returns:
378 None
379 """
380 # check if we performed the desired amount of sweeps
381 if S.status.iter == S.params.maxiter:
382 L = S.levels[0]
384 # compute next step size
385 order = S.status.iter # embedded error estimate is same order as time marching
387 e_est = self.get_local_error_estimate(controller, S)
388 L.status.dt_new = self.compute_optimal_step_size(
389 self.params.beta, L.params.dt, self.params.e_tol, e_est, order
390 )
391 self.log(f'Adjusting step size from {L.params.dt:.2e} to {L.status.dt_new:.2e}', S)
393 return None
395 def get_local_error_estimate(self, controller, S, **kwargs):
396 """
397 Get the embedded error estimate of the finest level of the step.
399 Args:
400 controller (pySDC.Controller): The controller
401 S (pySDC.Step): The current step
403 Returns:
404 float: Embedded error estimate
405 """
406 return S.levels[0].status.error_embedded_estimate
409class AdaptivityRK(Adaptivity):
410 """
411 Adaptivity for Runge-Kutta methods. Basically, we need to change the order in the step size update
412 """
414 def setup(self, controller, params, description, **kwargs):
415 defaults = {}
416 defaults['update_order'] = params.get('update_order', description['sweeper_class'].get_update_order())
417 return {**defaults, **super().setup(controller, params, description, **kwargs)}
419 def get_new_step_size(self, controller, S, **kwargs):
420 """
421 Determine a step size for the next step from an embedded estimate of the local error of the current step.
423 Args:
424 controller (pySDC.Controller): The controller
425 S (pySDC.Step): The current step
427 Returns:
428 None
429 """
430 # check if we performed the desired amount of sweeps
431 if S.status.iter == S.params.maxiter:
432 L = S.levels[0]
434 # compute next step size
435 order = self.params.update_order
436 e_est = self.get_local_error_estimate(controller, S)
437 L.status.dt_new = self.compute_optimal_step_size(
438 self.params.beta, L.params.dt, self.params.e_tol, e_est, order
439 )
440 self.log(f'Adjusting step size from {L.params.dt:.2e} to {L.status.dt_new:.2e}', S)
442 return None
445class AdaptivityResidual(AdaptivityBase):
446 """
447 Do adaptivity based on residual.
449 Since we don't know a correlation between the residual and the error (for nonlinear problems), we employ a simpler
450 rule to update the step size. Instead of giving a local tolerance that we try to hit as closely as possible, we set
451 two thresholds for the residual. When we exceed the upper one, we reduce the step size by a factor of 2 and if the
452 residual falls below the lower threshold, we double the step size.
453 Please setup these parameters as "e_tol" and "e_tol_low".
454 """
456 def setup(self, controller, params, description, **kwargs):
457 """
458 Define default parameters here.
460 Default parameters are:
461 - control_order (int): The order relative to other convergence controllers
462 - e_tol_low (float): Lower absolute threshold for the residual
463 - e_tol (float): Upper absolute threshold for the residual
464 - use_restol (bool): Restart if the residual tolerance was not reached
465 - max_restarts: Override maximum number of restarts
467 Args:
468 controller (pySDC.Controller): The controller
469 params (dict): The params passed for this specific convergence controller
470 description (dict): The description object used to instantiate the controller
472 Returns:
473 (dict): The updated params dictionary
474 """
475 defaults = {
476 "control_order": -45,
477 "e_tol_low": 0,
478 "e_tol": np.inf,
479 "use_restol": False,
480 "max_restarts": 99 if "e_tol_low" in params else None,
481 "allowed_modifications": ['increase', 'decrease'], # what we are allowed to do with the step size
482 }
483 return {**defaults, **params}
485 def setup_status_variables(self, controller, **kwargs):
486 """
487 Change maximum number of allowed restarts here.
489 Args:
490 controller (pySDC.Controller): The controller
492 Reutrns:
493 None
494 """
495 from pySDC.implementations.convergence_controller_classes.basic_restarting import BasicRestarting
497 if self.params.max_restarts is not None:
498 conv_controllers = controller.convergence_controllers
499 restart_cont = [me for me in conv_controllers if BasicRestarting in type(me).__bases__]
501 if len(restart_cont) == 0:
502 raise NotImplementedError("Please implement override of maximum number of restarts!")
504 restart_cont[0].params.max_restarts = self.params.max_restarts
505 return None
507 def check_parameters(self, controller, params, description, **kwargs):
508 """
509 Check whether parameters are compatible with whatever assumptions went into the step size functions etc.
511 Args:
512 controller (pySDC.Controller): The controller
513 params (dict): The params passed for this specific convergence controller
514 description (dict): The description object used to instantiate the controller
516 Returns:
517 bool: Whether the parameters are compatible
518 str: The error message
519 """
520 if controller.params.mssdc_jac:
521 return (
522 False,
523 "Adaptivity needs the same order on all steps, please activate Gauss-Seidel multistep mode!",
524 )
526 return True, ""
528 def get_new_step_size(self, controller, S, **kwargs):
529 """
530 Determine a step size for the next step.
531 If we exceed the absolute tolerance of the residual in either direction, we either double or halve the step
532 size.
534 Args:
535 controller (pySDC.Controller): The controller
536 S (pySDC.Step): The current step
538 Returns:
539 None
540 """
541 # check if we performed the desired amount of sweeps
542 if S.status.iter == S.params.maxiter:
543 L = S.levels[0]
545 res = self.get_local_error_estimate(controller, S)
547 dt_planned = L.status.dt_new if L.status.dt_new is not None else L.params.dt
549 if (
550 res > self.params.e_tol or (res > L.params.restol and self.params.use_restol)
551 ) and 'decrease' in self.params.allowed_modifications:
552 L.status.dt_new = min([dt_planned, L.params.dt / 2.0])
553 self.log(f'Adjusting step size from {L.params.dt:.2e} to {L.status.dt_new:.2e}', S)
554 elif res < self.params.e_tol_low and 'increase' in self.params.allowed_modifications:
555 L.status.dt_new = max([dt_planned, L.params.dt * 2.0])
556 self.log(f'Adjusting step size from {L.params.dt:.2e} to {L.status.dt_new:.2e}', S)
558 return None
560 def get_local_error_estimate(self, controller, S, **kwargs):
561 """
562 Get the residual of the finest level of the step.
564 Args:
565 controller (pySDC.Controller): The controller
566 S (pySDC.Step): The current step
568 Returns:
569 float: Embedded error estimate
570 """
571 return S.levels[0].status.residual
574class AdaptivityCollocation(AdaptivityForConvergedCollocationProblems):
575 """
576 Control the step size via a collocation based estimate of the local error.
577 The error estimate works by subtracting two solutions to collocation problems with different order. You can
578 interpolate between collocation methods as much as you want but the adaptive step size selection will always be
579 based on the last switch of quadrature.
580 """
582 def setup(self, controller, params, description, **kwargs):
583 """
584 Add a default value for control order to the parameters.
586 Args:
587 controller (pySDC.Controller): The controller
588 params (dict): Parameters for the convergence controller
589 description (dict): The description object used to instantiate the controller
591 Returns:
592 dict: Updated parameters
593 """
594 defaults = {
595 "adaptive_coll_params": {},
596 "num_colls": 0,
597 **super().setup(controller, params, description, **kwargs),
598 "control_order": 220,
599 }
601 for key in defaults['adaptive_coll_params'].keys():
602 if type(defaults['adaptive_coll_params'][key]) == list:
603 defaults['num_colls'] = max([defaults['num_colls'], len(defaults['adaptive_coll_params'][key])])
605 if defaults['restart_at_maxiter']:
606 defaults['maxiter'] = description['step_params'].get('maxiter', 99) * defaults['num_colls']
608 return defaults
610 def setup_status_variables(self, controller, **kwargs):
611 self.status = Status(['error', 'order'])
612 self.status.error = []
613 self.status.order = []
615 def reset_status_variables(self, controller, **kwargs):
616 self.setup_status_variables(controller, **kwargs)
618 def dependencies(self, controller, description, **kwargs):
619 """
620 Load the `EstimateEmbeddedErrorCollocation` convergence controller to estimate the local error by switching
621 between collocation problems between iterations.
623 Args:
624 controller (pySDC.Controller): The controller
625 description (dict): The description object used to instantiate the controller
626 """
627 from pySDC.implementations.convergence_controller_classes.estimate_embedded_error import (
628 EstimateEmbeddedErrorCollocation,
629 )
631 super().dependencies(controller, description, **kwargs)
633 params = {'adaptive_coll_params': self.params.adaptive_coll_params}
634 controller.add_convergence_controller(
635 EstimateEmbeddedErrorCollocation,
636 params=params,
637 description=description,
638 )
640 def get_convergence(self, controller, S, **kwargs):
641 return len(self.status.order) == self.params.num_colls
643 def get_local_error_estimate(self, controller, S, **kwargs):
644 """
645 Get the collocation based embedded error estimate.
647 Args:
648 controller (pySDC.Controller): The controller
649 S (pySDC.Step): The current step
651 Returns:
652 float: Embedded error estimate
653 """
654 if len(self.status.error) > 1:
655 return self.status.error[-1][1]
656 else:
657 return 0.0
659 def post_iteration_processing(self, controller, step, **kwargs):
660 """
661 Get the error estimate and its order if available.
663 Args:
664 controller (pySDC.Controller.controller): The controller
665 step (pySDC.Step.step): The current step
666 """
667 if step.status.done:
668 lvl = step.levels[0]
669 self.status.error += [lvl.status.error_embedded_estimate_collocation]
670 self.status.order += [lvl.sweep.coll.order]
672 def get_new_step_size(self, controller, S, **kwargs):
673 if len(self.status.order) == self.params.num_colls:
674 lvl = S.levels[0]
676 # compute next step size
677 order = (
678 min(self.status.order[-2::]) + 1
679 ) # local order of less accurate of the last two collocation problems
680 e_est = self.get_local_error_estimate(controller, S)
682 lvl.status.dt_new = self.compute_optimal_step_size(
683 self.params.beta, lvl.params.dt, self.params.e_tol, e_est, order
684 )
685 self.log(f'Adjusting step size from {lvl.params.dt:.2e} to {lvl.status.dt_new:.2e}', S)
687 def determine_restart(self, controller, S, **kwargs):
688 """
689 Check if the step wants to be restarted by comparing the estimate of the local error to a preset tolerance
691 Args:
692 controller (pySDC.Controller): The controller
693 S (pySDC.Step): The current step
695 Returns:
696 None
697 """
698 if len(self.status.order) == self.params.num_colls:
699 e_est = self.get_local_error_estimate(controller, S)
700 if e_est >= self.params.e_tol:
701 S.status.restart = True
702 self.log(f"Restarting: e={e_est:.2e} >= e_tol={self.params.e_tol:.2e}", S)
704 def check_parameters(self, controller, params, description, **kwargs):
705 """
706 Check whether parameters are compatible with whatever assumptions went into the step size functions etc.
707 For adaptivity, we need to know the order of the scheme.
709 Args:
710 controller (pySDC.Controller): The controller
711 params (dict): The params passed for this specific convergence controller
712 description (dict): The description object used to instantiate the controller
714 Returns:
715 bool: Whether the parameters are compatible
716 str: The error message
717 """
718 if "e_tol" not in params.keys():
719 return (
720 False,
721 "Adaptivity needs a local tolerance! Please pass `e_tol` to the parameters for this convergence controller!",
722 )
724 return True, ""
727class AdaptivityExtrapolationWithinQ(AdaptivityForConvergedCollocationProblems):
728 """
729 Class to compute time step size adaptively based on error estimate obtained from extrapolation within the quadrature
730 nodes.
732 This error estimate depends on solving the collocation problem exactly, so make sure you set a sufficient stopping criterion.
733 """
735 def setup(self, controller, params, description, **kwargs):
736 from pySDC.implementations.convergence_controller_classes.check_convergence import CheckConvergence
738 defaults = {
739 'high_Taylor_order': False,
740 **params,
741 }
743 self.check_convergence = CheckConvergence.check_convergence
744 return {**defaults, **super().setup(controller, params, description, **kwargs)}
746 def get_convergence(self, controller, S, **kwargs):
747 return self.check_convergence(S)
749 def dependencies(self, controller, description, **kwargs):
750 """
751 Load the error estimator.
753 Args:
754 controller (pySDC.Controller): The controller
755 description (dict): The description object used to instantiate the controller
757 Returns:
758 None
759 """
760 from pySDC.implementations.convergence_controller_classes.estimate_extrapolation_error import (
761 EstimateExtrapolationErrorWithinQ,
762 )
764 super().dependencies(controller, description, **kwargs)
766 controller.add_convergence_controller(
767 EstimateExtrapolationErrorWithinQ,
768 description=description,
769 params={'high_Taylor_order': self.params.high_Taylor_order},
770 )
771 return None
773 def get_new_step_size(self, controller, S, **kwargs):
774 """
775 Determine a step size for the next step from the error estimate.
777 Args:
778 controller (pySDC.Controller): The controller
779 S (pySDC.Step): The current step
781 Returns:
782 None
783 """
784 if self.get_convergence(controller, S, **kwargs):
785 L = S.levels[0]
787 # compute next step size
788 order = L.sweep.coll.num_nodes + 1 if self.params.high_Taylor_order else L.sweep.coll.num_nodes
790 e_est = self.get_local_error_estimate(controller, S)
791 L.status.dt_new = self.compute_optimal_step_size(
792 self.params.beta, L.params.dt, self.params.e_tol, e_est, order
793 )
795 self.log(
796 f'Error target: {self.params.e_tol:.2e}, error estimate: {e_est:.2e}, update_order: {order}',
797 S,
798 level=10,
799 )
800 self.log(f'Adjusting step size from {L.params.dt:.2e} to {L.status.dt_new:.2e}', S)
802 return None
804 def get_local_error_estimate(self, controller, S, **kwargs):
805 """
806 Get the embedded error estimate of the finest level of the step.
808 Args:
809 controller (pySDC.Controller): The controller
810 S (pySDC.Step): The current step
812 Returns:
813 float: Embedded error estimate
814 """
815 return S.levels[0].status.error_extrapolation_estimate
818class AdaptivityPolynomialError(AdaptivityForConvergedCollocationProblems):
819 """
820 Class to compute time step size adaptively based on error estimate obtained from interpolation within the quadrature
821 nodes.
823 This error estimate depends on solving the collocation problem exactly, so make sure you set a sufficient stopping criterion.
824 """
826 def setup(self, controller, params, description, **kwargs):
827 from pySDC.implementations.convergence_controller_classes.check_convergence import CheckConvergence
829 defaults = {
830 'control_order': -50,
831 **super().setup(controller, params, description, **kwargs),
832 **params,
833 }
835 self.check_convergence = CheckConvergence.check_convergence
836 return defaults
838 def get_convergence(self, controller, S, **kwargs):
839 return self.check_convergence(S)
841 def dependencies(self, controller, description, **kwargs):
842 """
843 Load the error estimator.
845 Args:
846 controller (pySDC.Controller): The controller
847 description (dict): The description object used to instantiate the controller
849 Returns:
850 None
851 """
852 from pySDC.implementations.convergence_controller_classes.estimate_polynomial_error import (
853 EstimatePolynomialError,
854 )
856 super().dependencies(controller, description, **kwargs)
858 controller.add_convergence_controller(
859 EstimatePolynomialError,
860 description=description,
861 params={},
862 )
863 return None
865 def get_new_step_size(self, controller, S, **kwargs):
866 """
867 Determine a step size for the next step from the error estimate.
869 Args:
870 controller (pySDC.Controller): The controller
871 S (pySDC.Step): The current step
873 Returns:
874 None
875 """
876 if self.get_convergence(controller, S, **kwargs):
877 L = S.levels[0]
879 # compute next step size
880 order = L.status.order_embedded_estimate
882 e_est = self.get_local_error_estimate(controller, S)
883 L.status.dt_new = self.compute_optimal_step_size(
884 self.params.beta, L.params.dt, self.params.e_tol, e_est, order
885 )
887 self.log(
888 f'Error target: {self.params.e_tol:.2e}, error estimate: {e_est:.2e}, update_order: {order}',
889 S,
890 level=10,
891 )
892 self.log(f'Adjusting step size from {L.params.dt:.2e} to {L.status.dt_new:.2e}', S)
894 return None
896 def get_local_error_estimate(self, controller, S, **kwargs):
897 """
898 Get the embedded error estimate of the finest level of the step.
900 Args:
901 controller (pySDC.Controller): The controller
902 S (pySDC.Step): The current step
904 Returns:
905 float: Embedded error estimate
906 """
907 return S.levels[0].status.error_embedded_estimate