Next: Individual-based Measures
Up: Cosmos Experiments 1:Detailed Analysis
Previous: Cosmos Experiments 1:Detailed Analysis
Analysis and Visualisation Techniques
One of the reasons that is often given for pursuing research with
artificial evolutionary systems is that they allow us to study
evolution in carefully controlled conditions, and to record as much
data as we wish during the course of an evolutionary run. However, the
very fact that we can record anything we wish to can sometimes lead to
problems. Without care, it is very easy to drown in megabytes of data. In
this section, we consider some of the questions we might want to ask
concerning the behaviour of an evolutionary system, and describe the
measures that were used in the Cosmos runs to provide answers to these
questions.
There is a broad distinction that can be made between measures which
relate to individual organisms within the population, and measures
which relate to the population as a whole. We will use both types when
analysing the Cosmos runs in this chapter.
Tim Taylor
1999-05-29