Ecosystem+based+problem+solving

Prev: Mutational priming

TODO List
 * REF: F de Boer Hogeweg (now published)
 * Include latest results
 * REF: Pagie Hogeweg
 * PICTURES!

=Ecosystem based problem solving=

So far we have been exploring the idea that space is important for: i.e. individual-based versus ecosystem-based (cf waves) **problem solving**. However, we have not looked at problem solving //together//.
 * clever individuals and population-based diversity
 * but also for population-based competition (group-selection?)

If we look at ecosystems (link) it is clear that there are specialistic individuals, but they cannot do things alone. So the question becomes: //is cooperative problem solving easy//?
 * in the hypercycle cooperative entities generate more information together, i.e. a cooperative solution which is population-based diversity
 * in the ligase system there was no evolution (i.e. predefined). //If we put it in, is it stable?//

Focusing on ecosystems:
 * most population models don't take **conservation of mass** into account.
 * and **just being there** is neglected, while it should be considered in terms of generating new niches through garbage, energy transduction, information transduction (processing of sunlight). //There is no magic coat to be invisible//.
 * probably all considerations of diversity in ecosystems are flawed because this is not take seriously.

In an artificial system however, we can phrase this problem in an interesting way, while in biology that is often not that clear. For this end Folkert de Boer and P. Hogeweg developed a function optimizing model in the following way:
 * cf. Pagie and Hogeweg (REF) they developed a LISP program which would solve a complex polynomial, where we as we have seen, populations can integrate information over generations.
 * here we ask: //can it evolve individual-based or population-based (co-operation) and what is more stable?//

The model:
 * in plane 1 we have the prey with  coordinates
 * there are predators which evolve LISP functions f(x,y) to solve the predefined polynomial F(x,y)
 * in plane 2 are the left-overs (detritus) from predation  = F(x,y) - f(x,y)
 * and there is a detritus eater: this evolves to match the left-overs g(x,y)
 * predators and detritus do not see each other and only detritus eater sees detritus

The results of this model show:
 * waves of high X values through a field of low X values: i.e. speciation in the prey
 * evolution proceeds: spreads out low X, high Y to low X, Y to the different corners of the function space.
 * this leads to very nice solutions
 * nice partitioning of polynomial between predator and detritus eater
 * first population-based diversity
 * then super beast takes over and detritus eater dies out
 * relative to information threshold: go for population-based solution!