142 ◾ Biological Computation
in practice the algorithm tests the tness of individuals and only indi-
rectly the tness of schemas; therefore, a small number of individuals
represent a larger number of possible schemas, and we have implicit
parallelism. On the other hand, if the population is too small, the
sampling of schemas it provides may be insucient and hence prob-
ably biased.
23.
a. e natural way is to represent solutions as binary chromosomes
of length 10. If a bit has value 1, the card belongs to the rst pile
and otherwise to the second pile. To compute the tness of a solu-
tion we will construct an error function and attempt to minimize
it. e function has to take into account both piles relative to
their target values ...