crack instabilities around tips in Molecular Modellings

Considering the MD (molecualr dynamics) simulation programs, they enable us to define the initial crack and then using different theories they propagate the crack. This process is actually a dynamic feature at least when the sample is going to fail. Here is the question that present in the most modellers assumptions, which will limit the simulation or maybe it is not possible to simulate the process with out these assumptions. One of them which I would like to know your ideas about is the linear velocity which come into conclusions before the simulations start. Actually is this linear velocity remains constant or increase with definite constant acceleration (rate) as crack propagates? I think the answer is No, so why we assume that this theory is accurate?

The second question is how we are sure or (how the program is sure!!) that the crack which has initiated will get into stability ( steady phase) while it is propagating in the effective direction through the sample? If we consider dislocations or diffusions mobility or even ductile to brittle transition( if we dont set initial temperature to be zero) then what can be the reply for this question?

Mahdi Kazemzadeh