With the start of the new year it is time to get the Nested Simulation Proposal for SED-ML ready for wide-spread adoption. I believe nested simulations are vital for SED-ML so that we can cover a much larger variety of simulation experiments. I think it is especially important NOT to create a new simulation class for every single different simulation we perform on a model. By just defining two simulation classes:
- One Step: which brings the model to the next desired output step.
- Nested Simulation: which allows running over another simulation task, while changing multiple models parameters with computed values from ranges.
it is possible to construct a large number of simulations that are currently carried out. I’ve taken these past weeks to fully flesh out all the details and the document is now available from Nature Proceedings:
The new version describes in detail all attributes and elements and features a number of examples (see below)
NOTE: This proposal only covers the generation of the data, not the visualization. In other words this proposal allows to generate n-dimensional data sets, while currently our DataGenerators can not access the values. I believe the two issues should be handled in different proposals.
Just a brief overview of the examples:
|Steady state scan|
|Pulsing a parameter during a simulation|
|Multiple Stochastic Traces|
|2D Steady State scan|
Just for completeness sake: here the link to the old version: