Track back in time

Summary

We want to track a beam given beam back in time throughout the lattice. This can either be a followup track or a given/generated distribution.

Motivation

From measurements or intuition you construct a final phase space P_{final}. One can then ask the question how would you generate P_{initial} and transport it. By tracking back (maybe using the optimiser) one can hope to achieve such a goal.

Approach (so far only for a straight machine ...)

1. The TRACK command is augment with Boolean TRACKBACK=
2. In ParallelTTracker the double member arrowOfTime_m is set to -1 in case of TRACKBACK=TRUE
3. In ParallelTTracker::invertP itsBunch->P *= arrowOfTime_m and itsBunch_m->RefPartP_m *= arrowOfTime_m; is set invertPz (need to update Doxygen)

This is as far as I have implemented it, if I remember how to do the branch/merge stuff :) I will commit.

To Do

0. All electrostatic elements do not need to be changed
1. Space charge needs to negate its action in all 3 dimensions.
2. RF
3. Make it more general i.e. 3. in Approach above