1036 Simulations from paper to investigate behavior of IDE model #1130

annawendler · 2024-09-27T07:09:46Z

Changes and Information

Add simulations and plots from paper to main branch
In particular, the following simulations can be conducted with the added scripts:
- Changepoint simulation
- Scenario based on the spread of COVID-19 in Germany
- Investigation of the rate of convergence of the implemented non-standard numerical solver

Merge Request - Guideline Checklist

Please check our git workflow. Use the draft feature if the Pull Request is not yet ready to review.

Checks by code author

Every addressed issue is linked (use the "Closes #ISSUE" keyword below)
New code adheres to coding guidelines
No large data files have been added (files should in sum not exceed 100 KB, avoid PDFs, Word docs, etc.)
Tests are added for new functionality and a local test run was successful (with and without OpenMP)
Appropriate documentation for new functionality has been added (Doxygen in the code and Markdown files if necessary)
Proper attention to licenses, especially no new third-party software with conflicting license has been added
(For ABM development) Checked benchmark results and ran and posted a local test above from before and after development to ensure performance is monitored.

Checks by code reviewer(s)

Corresponding issue(s) is/are linked and addressed
Code is clean of development artifacts (no deactivated or commented code lines, no debugging printouts, etc.)
Appropriate unit tests have been added, CI passes, code coverage and performance is acceptable (did not decrease)
No large data files added in the whole history of commits(files should in sum not exceed 100 KB, avoid PDFs, Word docs, etc.)
On merge, add 2-5 lines with the changes (main added features, changed items, or corrected bugs) to the merge-commit-message. This can be taken from the briefly-list-the-changes above (best case) or the separate commit messages (worst case).

Closes #1036
Closes #1097
Closes #411

…into 1036-ide-simulations

…e-of-the-discretization-scheme-from-the-ide-model' into 1036-ide-simulations

…nfirmedCasesNoAgeEntry and ConfirmedCasesDataEntry

…and ConfirmedCasesNoAgeEntry

…-both-confirmedcasesnoageentry-and-confirmedcasesdataentry' into 1036-ide-simulations

lenaploetzke

First part of review

lenaploetzke · 2025-01-23T08:29:43Z

...mulations/2024_Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/README.md

+
+- Section 5.1: With the file [ide_convergence_rate](ide_convergence_rate.cpp) simulations with an integro-differential-based model (IDE model) and an ordinary differential-based model (ODE model) with different time step sizes can be performed to analyze the order of convergence of the introduced nonstandard numerical scheme. The visualization of the results can be done with [plot_convergence](plot_convergence.py).
+
+- Section 5.2: In [ide_changepoints](ide_chanepoints.cpp) a simulation with both our IDE and an ODE model is performed where a change in the contact rate is applied after two days to investigate the model behavior at change points. The results can be visualized with [plot_changepoints](plot_changepoints.py).


Suggested change

- Section 5.2: In [ide_changepoints](ide_chanepoints.cpp) a simulation with both our IDE and an ODE model is performed where a change in the contact rate is applied after two days to investigate the model behavior at change points. The results can be visualized with [plot_changepoints](plot_changepoints.py).

- Section 5.2: In [ide_changepoints](ide_chanepoints.cpp), a simulation with both our IDE and an ODE model is performed, where a change in the contact rate is applied after two days to investigate the model behavior at change points. The results can be visualized with [plot_changepoints](plot_changepoints.py).

lenaploetzke · 2025-01-23T08:30:38Z

...mulations/2024_Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/README.md

+
+Below is an overview of the files and the paper sections they belong to. 
+
+- Section 5.1: With the file [ide_convergence_rate](ide_convergence_rate.cpp) simulations with an integro-differential-based model (IDE model) and an ordinary differential-based model (ODE model) with different time step sizes can be performed to analyze the order of convergence of the introduced nonstandard numerical scheme. The visualization of the results can be done with [plot_convergence](plot_convergence.py).


Suggested change

- Section 5.1: With the file [ide_convergence_rate](ide_convergence_rate.cpp) simulations with an integro-differential-based model (IDE model) and an ordinary differential-based model (ODE model) with different time step sizes can be performed to analyze the order of convergence of the introduced nonstandard numerical scheme. The visualization of the results can be done with [plot_convergence](plot_convergence.py).

- Section 5.1: With the file [ide_convergence_rate](ide_convergence_rate.cpp) simulations with an integro-differential equation-based model (IDE model) and an ordinary differential equation-based model (ODE model) with different time step sizes can be performed to analyze the order of convergence of the introduced nonstandard numerical scheme. The visualization of the results can be done with [plot_convergence](plot_convergence.py).

Maybe nonstandard numerical scheme for the IDE model? Maybe add also somewhere a SECIR- ?

lenaploetzke · 2025-01-23T08:37:49Z

...mulations/2024_Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/README.md

+
+- Section 5.3: The COVID-19 inspired scenario can be run with [ide_covid_inspired_scenario](ide_covid_inspired_scenario.cpp).
+
+    For the transition probabilities we use data from [Assessment of effective mitigation and prediction of the spread of SARS-CoV-2 in Germany using demographic information and spatial resolution](https://doi.org/10.1016/j.mbs.2021.108648) and [Covasim: An agent-based model of COVID-19 dynamics and interventions](https://doi.org/10.1371/journal.pcbi.1009149). There, the data is age-resolved. We consider a model without age resolution which is why compute according parameters by averaging over the age groups. This is done in [compute_parameters](compute_parameters.cpp).


Isn't this also the case for the cangepoints? Maybe you can reformat this somehow or add a comment

lenaploetzke · 2025-01-23T08:56:52Z

...mulations/2024_Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/README.md

+
+    For the transition probabilities we use data from [Assessment of effective mitigation and prediction of the spread of SARS-CoV-2 in Germany using demographic information and spatial resolution](https://doi.org/10.1016/j.mbs.2021.108648) and [Covasim: An agent-based model of COVID-19 dynamics and interventions](https://doi.org/10.1371/journal.pcbi.1009149). There, the data is age-resolved. We consider a model without age resolution which is why compute according parameters by averaging over the age groups. This is done in [compute_parameters](compute_parameters.cpp).
+
+    For the stay time distributions in the respective compartments we used lognormal distributions as in [Covasim: An agent-based model of COVID-19 dynamics and interventions](https://doi.org/10.1371/journal.pcbi.1009149). There, the mean and the standard deviation were given. In order to use it with our implementation, we computed the corresponding shape and scale parameters with [get_lognormal_parameters](get_lognormal_parameters.py).


Suggested change

For the stay time distributions in the respective compartments we used lognormal distributions as in [Covasim: An agent-based model of COVID-19 dynamics and interventions](https://doi.org/10.1371/journal.pcbi.1009149). There, the mean and the standard deviation were given. In order to use it with our implementation, we computed the corresponding shape and scale parameters with [get_lognormal_parameters](get_lognormal_parameters.py).

For the stay time distributions in the respective compartments, we used lognormal distributions as in [Covasim: An agent-based model of COVID-19 dynamics and interventions](https://doi.org/10.1371/journal.pcbi.1009149). There, the mean and the standard deviation were given. In order to use it with our implementation, we computed the corresponding shape and scale parameters with [get_lognormal_parameters](get_lognormal_parameters.py).

lenaploetzke · 2025-01-23T08:59:00Z

...mulations/2024_Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/README.md

+    - Data on reported cases can be downloaded with [getCaseData](../../../pycode/memilio-epidata/memilio/epidata/getCaseData.py). To obtain the data, use the options moving_average = 7 and impute_dates = True when downloading the data. For the other function parameters, one can use default parameters from [defaultDict](../../../pycode/memilio-epidata/memilio/epidata/defaultDict.py).
+
+    When simulating, we set the contact rate such that the daily new transmissions of the IDE simulation are consistent with the reported data at the beginning of the simulation. The contact scaling and the visualization is done in [run_and_plot_covid_inspired_scenario](run_and_plot_covid_inspired_scenario.py). For the scaling and the subsequent visualization we used case data from RKI (not averaged) and data on ICU patients as reported from DIVI.
+    - To obtain the case data, use [getCaseData](../../../pycode/memilio-epidata/memilio/epidata/getCaseData.py) with the option impute_dates = True when downloading.


Maybe add a comment that ma7 is not set here.

lenaploetzke · 2025-01-23T10:31:51Z

...Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/ide_convergence_rate.cpp

+
+    mio::osecir::Model<ScalarType> model_ode(1);
+
+    // Set initial values for compartments.


Maybe add comment that they are random

lenaploetzke · 2025-01-23T10:32:08Z

...Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/ide_convergence_rate.cpp

+
+    // Set parameters.
+    ScalarType cont_freq = simulation_parameter["cont_freq"];
+    // Set Seasonality=0 so that cont_freq_eff is equal to contact_matrix.


lenaploetzke · 2025-01-23T10:32:51Z

...Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/ide_convergence_rate.cpp

+                                                    simulation_parameter["total_population"]);
+
+    // Set parameters.
+    ScalarType cont_freq = simulation_parameter["cont_freq"];


why this extra variable

lenaploetzke · 2025-01-23T10:35:54Z

...Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/ide_convergence_rate.cpp

+        mio::isecir::Model model_ide(std::move(init_transitions), total_population, deaths, num_agegroups);
+
+        // Set parameters.
+        // Contact matrix; contact_matrix was already defined for ODE.


Suggested change

// Contact matrix; contact_matrix was already defined for ODE.

// contact_matrix was already defined for ODE.

lenaploetzke · 2025-01-23T12:29:54Z

...Wendler_Nonstandard_numerical_scheme_for_integro-differential_model/ide_convergence_rate.cpp

+
+    if (!result_dir.empty() && save_exponent > 0) {
+        // Create result directory if not existent yet.
+        boost::filesystem::path res_dir(result_dir);


boost filesystem header ?

lenaploetzke

really really nice work :)

lenaploetzke · 2025-01-23T14:39:01Z