I’m trying to implement a solver that will stop if one of the variables passes a certain value instead of time-stepping (I may thus change the title of this topic). Say e.g.
u < 0 || u > R for maximum range
There are two methods I’ve been looking at, but haven’t gotten anything working (out of lack of understanding or whatnot).
Q1: The first way I figured was probably possible was to use the
isoutofdomain input argument as
isoutofdomain = MyOutOfDomain(u, p, t). Is this correct? When I try it, the solver doesn’t finish (and I’ve solved without this input argument and it solves very quickly, so I’m quite confused as to how adding this would increase computation time by tonnes).
Q2: The second way I could do this is via a callback, i.e.
callback = cbRange with
cbRange = ContinuousCallback(RangeCond, RangeAffect!). But I’m not quite sure how to implement
RangeAffect! so that the solver stops.
Q3: How can I change the stopping condition from the default
tspan = (0, T) to stopping with one of the variables (say
u) reaching certain values instead? I understand that this means it is possible for this variable potentially not reaching its stopping range values, so there would need to be some sort of implementation to catch when this happens as well.
I figured this is the most minimal way to ask my question(s), but I can expand on the context for you if you want.
Thanks in advance!