From deae4025b86e8e3d58505b07feb24230523b8744 Mon Sep 17 00:00:00 2001 From: Ishaan Desai Date: Mon, 16 Sep 2024 16:08:39 +0200 Subject: [PATCH 1/5] Add info about adaptivity tuning parameters --- docs/configuration.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/docs/configuration.md b/docs/configuration.md index e7fe0b7..afeb006 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -121,6 +121,12 @@ Parameter | Description `every_implicit_iteration` | If True, adaptivity is calculated in every implicit iteration.
If False, adaptivity is calculated once at the start of the time window and then reused in every implicit time iteration. `similarity_measure`| Similarity measure to be used for adaptivity. Can be either `L1`, `L2`, `L1rel` or `L2rel`. By default, `L1` is used. The `rel` variants calculate the respective relative norms. This parameter is *optional*. +The primary tuning parameters for adaptivity are the history parameter $$ \Lambda $$, the coarsening constant $$ C_c $$, and the $$ C_r $$. Their effects can be interpreted as: + +- Higher values of the history parameter $$ \Lambda $$ imply lower significance of the adaptivity state in the previous timestep on the state in the current timestep. +- Higher values of the coarsening constant $$ C_c $$ imply that more active simulations from the previous timestep will remain active in the current timestep. +- Higher values of the refining constant $$ C_r $$ imply that less inactive points from the previous timestep will become active in the current timestep. + Example of adaptivity configuration is ```json From 18c7c1e29346c601dcdb34e438367b6d819db305 Mon Sep 17 00:00:00 2001 From: Ishaan Desai Date: Mon, 16 Sep 2024 16:36:15 +0200 Subject: [PATCH 2/5] Add Changelog entry --- CHANGELOG.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/CHANGELOG.md b/CHANGELOG.md index 3881626..447b6d2 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,5 +1,9 @@ # Micro Manager changelog +## latest + +- Add information about adaptivity tuning parameters https://github.com/precice/micro-manager/pull/131 + ## v0.5.0 - Use absolute values to calculate normalizing factor for relative norms in adaptivity https://github.com/precice/micro-manager/pull/125 From b5d57d92df3ba3934f15a4a70cd21f3027165d2b Mon Sep 17 00:00:00 2001 From: Ishaan Desai Date: Tue, 17 Sep 2024 10:51:05 +0200 Subject: [PATCH 3/5] Add keyword of tuning parameter Co-authored-by: mathiskelm <114579716+mathiskelm@users.noreply.github.com> --- docs/configuration.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/configuration.md b/docs/configuration.md index afeb006..c65d1fa 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -121,7 +121,7 @@ Parameter | Description `every_implicit_iteration` | If True, adaptivity is calculated in every implicit iteration.
If False, adaptivity is calculated once at the start of the time window and then reused in every implicit time iteration. `similarity_measure`| Similarity measure to be used for adaptivity. Can be either `L1`, `L2`, `L1rel` or `L2rel`. By default, `L1` is used. The `rel` variants calculate the respective relative norms. This parameter is *optional*. -The primary tuning parameters for adaptivity are the history parameter $$ \Lambda $$, the coarsening constant $$ C_c $$, and the $$ C_r $$. Their effects can be interpreted as: +The primary tuning parameters for adaptivity are the history parameter $$ \Lambda $$, the coarsening constant $$ C_c $$, and the refining constant $$ C_r $$. Their effects can be interpreted as: - Higher values of the history parameter $$ \Lambda $$ imply lower significance of the adaptivity state in the previous timestep on the state in the current timestep. - Higher values of the coarsening constant $$ C_c $$ imply that more active simulations from the previous timestep will remain active in the current timestep. From 7dd07509d175b38c898479ca2d755ecca199bd01 Mon Sep 17 00:00:00 2001 From: Ishaan Desai Date: Tue, 17 Sep 2024 10:53:55 +0200 Subject: [PATCH 4/5] Refer to similarity measures instead of adaptivity states while explaining the history parameter Co-authored-by: mathiskelm <114579716+mathiskelm@users.noreply.github.com> --- docs/configuration.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/configuration.md b/docs/configuration.md index c65d1fa..2a8c965 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -123,7 +123,7 @@ Parameter | Description The primary tuning parameters for adaptivity are the history parameter $$ \Lambda $$, the coarsening constant $$ C_c $$, and the refining constant $$ C_r $$. Their effects can be interpreted as: -- Higher values of the history parameter $$ \Lambda $$ imply lower significance of the adaptivity state in the previous timestep on the state in the current timestep. +- Higher values of the history parameter $$ \Lambda $$ imply lower significance of the similarity measures in the previous timestep on the similarity measure and thus adaptivity state in the current timestep. - Higher values of the coarsening constant $$ C_c $$ imply that more active simulations from the previous timestep will remain active in the current timestep. - Higher values of the refining constant $$ C_r $$ imply that less inactive points from the previous timestep will become active in the current timestep. From 1030135201c7c9feb732c8dbf66d01a3969f9e08 Mon Sep 17 00:00:00 2001 From: Ishaan Desai Date: Thu, 10 Oct 2024 15:06:37 +0200 Subject: [PATCH 5/5] Add practical bound recommendations for the coarsening and refining constants --- docs/configuration.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/configuration.md b/docs/configuration.md index 2a8c965..aedbed1 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -116,8 +116,8 @@ Parameter | Description `type` | Set to either `local` or `global`. The type of adaptivity matters when the Micro Manager is run in parallel. `local` means comparing micro simulations within a local partitioned domain for similarity. `global` means comparing micro simulations from all partitions, so over the entire domain. `data` | List of names of data which are to be used to calculate if micro-simulations are similar or not. For example `["temperature", "porosity"]`. `history_param` | History parameter $$ \Lambda $$, set as $$ \Lambda >= 0 $$. -`coarsening_constant` | Coarsening constant $$ C_c $$, set as $$ C_c < 1 $$. -`refining_constant` | Refining constant $$ C_r $$, set as $$ C_r >= 0 $$. +`coarsening_constant` | Coarsening constant $$ C_c $$, set as $$ 0 =< C_c < 1 $$. +`refining_constant` | Refining constant $$ C_r $$, set as $$ 0 =< C_r < 1 $$. `every_implicit_iteration` | If True, adaptivity is calculated in every implicit iteration.
If False, adaptivity is calculated once at the start of the time window and then reused in every implicit time iteration. `similarity_measure`| Similarity measure to be used for adaptivity. Can be either `L1`, `L2`, `L1rel` or `L2rel`. By default, `L1` is used. The `rel` variants calculate the respective relative norms. This parameter is *optional*.