I’m working with AMMR v1.0, especially the standing model. I found a strange distribution of load between the agonist and antagonist muscles while the model is in standing posture. You can see that the antagonist and agonist muscles are both activated. Moreover, some muscle fascicles with smaller moment arms gain more load from the opt. function in comparison with those with larger moment arms!
These two facts are in contrast with the optimization process. Please let us know what happens in your solver.
First of all, AMMR v1.0 is very old. I suggest you request an update to have access to more recent models.
Re your question, actually, it depends on the function of the muscles. If the muscles are uni-articular, recruiting both agonist and antagonist muscles is against effort minimization. Therefore, a muscle recruitment solver will only activate each at a time. However, if an agonist muscle spans two or more joints [or degrees of freedom (dof) to be more accurate], it might be activated at the same time that antagonist muscles (at one of those dof the multi-articular muscle spans) are on, which is due to the equilibrium requirement.
Although sufficiently bio-fidelic models have 3D muscle lines of action that span multiple joints, in a simplified case, if you have a 2D model with uni-articular muscles, there will not be any antagonistic coactivation.
Hope it helps,
Mohammad S. Shourijeh, PhD