Alastair Channon

Alastair Channon is Lecturer and part of the Natural Computation Group at the Schoool of Computer Science, University of Birmingham

Designs from artificial natures
Many artefacts, from sheepdogs to prescription drugs, are less the product of manual design than refinements of artefacts discovered in
nature. Can we construct "artificial natures" that generate many novel artefacts/designs from which we can harvest and refine useful ones?

Areas to be explored include:

• what classes of natural computation can be used to generate many novel artefacts/designs? (Natural selection based evolution, others?)
• what classes of systems will be likely to produce potentially useful artefacts/designs? (Must they be similar to our world / the target environment?)
• what classes of systems will be likely to produce artefacts/designs that a designer/selector can easily recognise as useful when they emerge?
• should we first harvest and then refine, or should refining a design (eg. through artificial selection) be part of the harvesting process?
• what applications can we envisage? In what realms? (Computer games, animation, other simulation based engineering, art, ...)

---