Abstract
Ecological, environmental and geophysical time series consistently exhibitthe characteristics of coloured (1/f^\b{eta}) noise. Here we briefly survey theliterature on coloured noise, population persistence and related evolutionarydynamics, before introducing coloured noise as an appropriate model forenvironmental variation in artificial evolutionary systems. To illustrate andexplore the effects of different noise colours, a simple evolutionary modelthat examines the trade-off between specialism and generalism in fluctuatingenvironments is applied. The results of the model clearly demonstrate a needfor greater generalism as environmental variability becomes `whiter', whilstspecialisation is favoured as environmental variability becomes `redder'. Pinknoise, sitting midway between white and red noise, is shown to be the point atwhich the pressures for generalism and specialism balance, providing someinsight in to why `pinker' noise is increasingly being seen as an appropriatemodel of typical environmental variability. We go on to discuss how the resultspresented here feed in to a wider discussion on evolutionary responses tofluctuating environments. Ultimately we argue that Artificial Life as a fieldshould embrace the use of coloured noise to produce models of environmentalvariability.