Base survival probability of charge carrier on time step length

This addresses one discussion of #196 (closed):

I am not sure how to improve this (or whether it needs improvement, actually), but I'm wondering about whether using the total time here is a good thing. I'm taking it to the extreme with an example:

Imagine a charge carrier that was produced in an undepleted region of the sensor. It diffuses around quite a bit and after, say 50 ns it reaches an electric field. During its drift it passes through a very small region with a low lifetime. It is inside this region for maybe 50 ps, but the time input used to decide whether it survives or not would be 50.something ns.

The above example in my eyes is unphysical, as the carrier shouldn't be aware of it being 50 ns old.

Thus I based the survival probability of charge carriers on the duration of a time step, not on the carrier's age. The following graph shows the result of a Toy MC with the current implementation and the proposed one. It starts with a lifetime of 1000, after a time of 500 it is reduced to 100. You can see that due to their age, in the current implementation the carriers recombine immediately, which is mitigated with the new one. Also it is independent of the length of the time step.

Probably the most tests will fail due to updated PRN usage - I would fix it if we agree on this approach.