The interdisciplinary working group on "Evolutionary
Transitions" was founded on May 6, 2003 at the Center Leo Apostel to study
the emergence of multiple levels of complexity, especially in biological,
social and computational systems.
A number of these wholes can in a later stage again be joined together, forming a supersystem of an even higher order. This produces a "dynamical hierarchy", i.e. a dynamical system exhibiting multiple levels of nested subcomponent structures. The components at each level, their properties, and the modes of interaction between them, all result from the ongoing interactions between the components of lower levels.
Subsequent transitions explain the fundamentally hierarchical evolution of complex systems, and indicate the general trend of increasing complexity, organisation and synergy that characterises evolution.
This subject ties in with:
The group is associated with the interfaculty
Center
Leo Apostel (CLEA) of the VUB, but welcomes members from any background
or affiliation interested in participating in its (provisionally Brussels-based)
activities. All people interested in joining are hereby invited. For more
info, contact F. Heylighen.
May 13: Francis Heylighen:
Various explanations have been proposed to explain how cooperation can nevertheless evolve: e.g. kin selection, reciprocal altruism, interiorised morals, market mechanisms, or state institutions. In this seminar I will propose the new abstract mechanism of aspect co-evolution to integrate all these as special cases. The main idea is that the components (cells, individuals, ...) that make up a collective are initially all similar, with similar needs and similar outputs. They thus will be in competition for the same resources, making it difficult to evolve cooperation. Two very different systems, though, can more easily enter into a mutualist relationship where the products of the one may be used as the resources of the other, and vice versa. E.g. a lichen is a symbiotic organism consisting of an alga and a fungus.
A similar complementary relationship can also
be attained within a collective, not so much between its components or
subsystems, but between functional levels of organization or "aspect systems",
that represent different relationships between the same physical components.
For example, markets and religions are different "aspect systems" of the
same system of society. Being dissimilar and relatively autonomous, while
interacting through their shared components, puts the aspect systems in
a relation of co-evolution, where they can mutually adapt to reach a symbiotic
configuration. This configuration can then act as a control mechanism that
suppresses free riders, thus safeguarding the cooperation.
The general model which can be derived from these phenomena serves as a collection of guidelines which can be used to design algorithms which produce composite solutions from simple ones. Yet, it is difficult in that context to build these kind of algorithms. It is often not clear in many cases how a task can be divided and what the influence is of the different components on each other (interaction). Hence, the most adequate decomposition should be learned automatically instead of being manually constructed.
Thus from a computer science perspective, instance of evolutionary transitions provide information on how to construct algorithms which are capable of both finding optimal solutions for a particular problem and doing this by hierarchically constructing these solution. There are different issues which need further investigation since it is often not trivial to translate biological phenomena to usable algorithms. One of the most important is whether these constuctions are static (or structural) hierarchies or whether they are dynamic hierarchies where a (stable) equilibrium has been reached, within and between levels.
This talk will address the computer science perspective on this issue
and discuss some initial steps which have been taken in designing an evolutionary
algorithm that, in the long run, is capable of transitions in complexity.
1. The history of system dynamics (especially the World3 model, used by the Club of Rome). This component will be important in order to evaluate the implicit goals of system dynamics in the past (‘70s and ‘80s).
2. The search for the social theoretic assumptions underlying system dynamics. For instance, David Lane from the London School of Economics has stated that Giddens’s structuration theory could be a good theoretical foundation of system dynamics. I disagree. I prefer a link with the theoretical research concerning complexity.
3. The difference between the quantitative and the qualitative approach of system dynamics. What are the advantages and disadvantages of both approaches ?
4. The link between uncertainty and system dynamics. Given different kinds of uncertainty, what can our system dynamics models say ? What can they do ?
5. The link between system dynamics and sustainable development.