Week 10 - Nov.4th

TODO for this week

1. fix Latex Reflection and Simulation from previous meeting

2. add detailed calculation part for triangle reflection

3. rootSolve, multiroot should work under assumption that we always have equation for the boundaries, which we do.

later we can try when incoming vector is not straight line

To Discuss

• Escape time by define tube length/width
• Previous bug: use current boundary to check, round 15 digits in single_reflect b/c sqrt calculations (works fine for triangle and bar mode). use current boundary to check as well

Discussed during Meeting

• when comparing =0, use abs(value) <tolerance to avoid rounding - to do
• Test in circle mode when entry pt is very close to left end. It should hit left boundary then h, but the left distance is so small that it's rounded to 0 so left boundary is ignored.
  • when solutions are <=0, assign Inf to t. However the t is not actually 0 but rounded to 0, thus we assign t_l = Inf.
  • R rounded c_l and c_r to 0 with no clear reason, we worked on this with digits = 22 (max) and it still does not work.
  • Solution print error and stop program when this happens. We will run 1e6 simulations and see how many fall under situation described above
• Discussed rootSolve. will implement. To be discussed: start pt (first guess) for rootSolve, can we provide a good guess