Set tolerances for 'incorrect' algorithms, when more strict than defaults

[luciansmith]

Contact Details

No response

Location

runBiosimulations

What happened?

Models that succeed using LSODA sometimes do not succeed using CVODE, due to needing to turn on/off the 'stiff' solver, and/or set tolerances to be more restrictive.

However, you currently can't create a SED-ML file that tells you what the tolerances should be for both algorithms. This should also be solved, but in the meantime, it should be possible to create SED-ML for LSODA with extra-strict tolerances which will work for both LSODA and CVODE.

However, currently, if the base algorithm is LSODA but the simulator uses CVODE, all algorithm parameters are ignored.

Often this is a good thing: LSODA tolerances are often set higher than default CVODE tolerances, and you need the defaults to successfully simulate the model. But in the reverse case, when the tolerances are more restrictive than the defaults, biosimulations should use the more restrictive values.

What made it happen?

Simulate a LSODA simulation on a CVODE simulator

Version

Production (Deployed on a .org domain)

What browsers are you seeing the problem on?

No response

Relevant log output

No response

Code of Conduct

• I agree to follow this project's Code of Conduct