Add option for survival biasing source normalization

[yrrepy]

See #2673 for history.
Re-pull-requesting here as the code is re-synced with the latest dev branch.
@eepeterson

I believe most things from the original PR are addressed, though I have left the ability to switch it off (and thus a setting for).

[paulromano]

Thanks @yrrepy and sorry this one has lingered for a long time. I made a few updates here as did @pshriwise. Note that I changed the attribute wgt0 to wgt_born to be consistent with other terminology in the code where we're referring to an attribute at the time a source particle was born.

I'm good with this now but would like to get approval from @pshriwise and @eepeterson too.

[pshriwise]

Based on the current implementation, if we stick with it we probably want to update this language from "history" to "particle".

[pshriwise]

FWIW I worked with this a little bit today using a test model that was a 10cmx10cmx10cm tungsten block (artificially high density) with a planar neutron source on the lower x/y plane, mono-directional in (0, 0, 1), and monoenergetic (14.1 MeV).

I applied a weight window generator (using MAGIC) on a 100 x 100 x 100 mesh and here's what I saw:

Analog (No ww)	Survival Biasing On

Survival Biasing Off	Survival Biasing Normalization

Note that this is with the Particle::weight_born attribute set in the Particle::from_source method. Otherwise I saw no effect. A material with more absorption would probably be a better test case, I can look into that as well if needed. But it at least seems that this method allows for implicit capture with weight windows applied without negating the weight windows severely.

It would be good to test this for source weights as well as you said @yrrepy. Some sources of varying strengths with uniform_source_sampling turned on might be interesting to look at as well.