Improve bootstrap process in Stock Synthesis

[k-doering-NOAA]

Imported from redmine, Issue #72710 Opened by @k-doering-NOAA on 2019-12-19 Status when imported: New

This was a wishlist item for a future development feature - full description is "improve the bootstrap process to more fully mimic the sampling process rather than just add residuals to the expected value of observations."

@richard.methot, feel free to add details and change the priority.

[k-doering-NOAA]

comment from @RickMethot on 2020-02-28: for example: currently the multinomial sample of expected age' composition is drawn after the ageerror matrix has done its blurring. This is not best. Better to draw sample from the true sampled agecomp, and then apply age_error. Same can be done for generating the mean length at age'