Choosing between these three types of models

It is important to say that none of the modelling frameworks mentioned above (or any that we did not introduce) is generally superior to the others. It always depends on the research question and the data available to validate the model. 

If the research question is purely qualitative (e.g. is the presence of A needed for B?), Boolean network models might be a good choice. If the research question is quantitative (e.g. how much of A do you need to activate B?), Boolean network models will not be able to answer the question and difference or differential equation models would be a better choice.

And similarly, if the data that you want to validate your model behavior or predictions with is qualitative (e.g. presence or absence of components), Boolean network modelling might be a good framework. Quantitative modelling might not be a good choice in that case because you are missing information to validate the model’s dynamics. However, if you have quantitative data, difference or differential equation models could be a good choice. Whether you measured your data at certain time points or continuously might influence the choice between difference and differential equation models. But it is important to note that it is possible to validate differential equation models with data that was measured in discrete time steps.

In practice, the choice of modelling framework often depends on real-life constraints. In addition to available data and the research question, it might be influenced by a researcher’s previous exposure to a certain framework and knowledge.