next_inactive up previous
Up: High Performance Coders

Evolutionary systems


I am principally interested the properties of evolutionary systems, in particular what properties are needed for evolutionary systems to produce unbounded novelty. I believe that all creative processes - eg biological evolution, human economic systems and human thought are all fundamentally evolutionary, that is are created through variation and selection.

Although the basic outline of Darwinian evolution can be implemented in a computer program, to date, none of these systems have shown unbounded creativity. This, in fact, is one of the open problems in artificial life. In many cases, the reason is that evolutionary algorithms are used to optimise something, encoded as a fitness function -- when fitness is optimised, evolution ceases. Natural evolutionary systems such as the biosphere, or economic systems do not have a well defined fitness function (in spite of what many biologists might say). Adaption does accur, but what the components of the system are adapting to varies as the other components adapt, something called coevolution.

However, this is not the whole story, as no artificial coevolutionary system (such as Tierra) has exhibited unbounded creativity either. In fact, even the issue of just how to measure creativity in processes has only received answers in the last 5 years or so.

Agent Based Modelling

In more recent years, I have entered the world of software for agent based modelling, as the EcoLab software I have been developing continuously since 1993 became progressively more general purpose.

Ontological Theories of Everything

See Theory of Nothing

next_inactive up previous
Up: High Performance Coders
Russell Standish 2016-03-05