
Develop an estimate of the size of the arrays that will hold the
current and next generation. Using the mutation rate and the population
size you can determine the expected number of new mutations per
generation, and from there, using an estimate of the number of
generations to extinction [see the table in the following
exercise], you can derive an upper bound for the
length of the genome.

Modify your asexual simulator to account for sexual reproduction,
using the techniques outlined in this section. Compare the output
of the new simulator with selected results from the following
table:

Modify the sexual simulator so it models a dioecious mating
system, i.e. classify all individuals as either males or females,
and draw one parent from each class. As a further variation,
you might want to simulate monogamy: at the beginning of the
build_next_generation
procedure, form pairs of males and
females, and draw pairs at random to create new individuals.

For the procedure that performs recombination based on
choosing randomly from a parent's two genes,
implement and test the random bit generator described in
Numerical Recipes [3].
See the CSEP chapter on random
number generators for tests of independence and other metrics
for measuring the quality of random number generators.
How does a random bit generator compare to the sequence of
bits taken from a random integer generator? Is there a
hidden bias in the random integer generator that would make
it choose nonrandom segments from one strand or the other
in the parent?