Next: Method
Up: Cosmos Experiments 2: Exploring
Previous: Cosmos Experiments 2: Exploring
Some of the results reported in the previous chapter suggest that
chance events may have played a significant role in the outcome
of the run (see Section 5.2.7). As discussed in
Section 2.3.5, it is generally accepted that
contingency (``historical accident'') has played a large part in
determining the course of evolution on Earth, although there is still
much debate as to the relative importance of chance events versus
other factors (e.g. the inherent self-organisational properties of the
biochemical world, discussed in Section 2.1.1).
With artificial evolutionary systems we have the advantage of being
able to ``replay'' evolution under experimental control. When
considering the performance of an evolutionary system, we
generally wish to disentangle the relative influence of three factors:
(1) contingency, (2) performance due to particular details of the
system's design, and (3) performance which may be general to a wide class of
evolutionary systems [Taylor & Hallam 97]. In this section, an experiment
is reported in which Cosmos was run a number of times, varying just
the random number seed between runs. The results will give us a better
idea of the role of contingency in the system.
Next: Method
Up: Cosmos Experiments 2: Exploring
Previous: Cosmos Experiments 2: Exploring
Tim Taylor
1999-05-29