Hi,
I am sorry that I missed this one a while ago when I tried to fix
infiite loops in SimTrack due to some other reason, minor steps were not
bound from below in original codes so it shrinks to some unbelievably
small values. I fixed that one but did not pay attention to this.
Chi
On Fri, 25 Apr 2003, Chris Crawford wrote:
> hi,
> here are some details of the inifinite loop in TBLSimTrack. i fixed
> in in version '-r 1.11.2.16'. the runga kutta is divided into major
> steps of 10cm. during each major step, the step-size is decreased in an
> inner loop until the precision is sufficient, and at the end of this
> loop, the major step is registered and minor steps are reset.
> however, two different tests are used to determine when to reset, and
> when to register the old step:
>
> while(totalLen < step-.000001) {
> // need to prevent h from decaying to zero
> h2 = h/2.0;
> h4 = h/4.0;
> ....
>
> totalLen += h;
> if (totalLen > step-.0000001) { //&& iflag == 0
> fXx[++fStepN] = x1; // increment fStep by 1
> ... }
>
> this gives the possibility of reseting the step without moving on to
> the next one. if anyone sees a reason for doing it this way, i would
> appreciate hearing from you.
> --thanks, chris
>
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:29 EST