How+to+model+complex+systems+(Major+transitions)

Prev: Defining properties of complex systems Next: Ecosystem complexity, selection and insight

TODO List
 * REF: major transitions
 * REF BOIDS
 * REF Beltman Maree

=How to model complex systems?=


 * how do biological systems become complex?
 * certainly not due to survival of the fittest mechanism: not all systems evolve to complexity, many go to simpler!
 * often it is ignored: wondering where complexity comes from!

Maynard Smith noted that all of evolutionary theory does not help to understand how particular properties evolve (cf the elephant problem). Together with Szathmary he published //The Major Transitions in Evolution// (1995) to try and identify real major changes in information processing, or the map from the prebiotic soup to now. Here we compare their approach and interpretation relative to ours:

to **indvidual replicators** because if they do not use this constraint they could make any story they want! || - because of the formation of higher-level entities, both premises are **not true**: -- (a) fitness is time dependent on different timescales -- (b) not only in terms of individual replicators, but spirals || can replicate only as part of larger whole afterwards || - have observed this: in spirals, sub-functionalization of duplicated genes: cannot do it alone, losing abilitiy due to being in environment where something done by another who certainly will do it! Thus speciation, division of labour in humans and insects. || conflicting selection pressures at different levels || - we can also see multi-level selection without conflict! e.g. host-parasite wave spirals. If in a chaotic region, evolution is such that chaotic region is formed and vice-versa. Thus we have self-reinforcing spatial patterns. - also of interest here is population versus individual-based diversity: in plasmids, the individual level selection is in the same direction to higher level ||
 * **They did** || **We did** ||
 * - reconstruct: what did happen (at some times flawed) || - assume: Darwinian selection and local interactions (quite sure of these) ||
 * - in light of: Darwinian selection + chemistry ||  ||
 * - infer intermediate steps + common principles || - observe what happens due to those assumptions: what DOES happen ||
 * = **Bottom line for both:** ||
 * = **multi-level evolution (entities first on own become part of larger system)** ||
 * - must be explained in terms of **immediate selective advantage**
 * - transitions are: entities capable of independent replication before transition
 * - if there are transition to higher-levels: it is exception to the rule and there are

Thus we come up with the following overview:
 * **Local interactions and Darwinian evolution:** feedback from emerging pattern to local rules self-reinforce
 * **Local interactions and intertwined multiple space and time scales:** competition depends on global interaction structure
 * **Local interactions is important for the evolution of complex replicators**: this suggestion needs to be pinpointed, changes circumstances on all kinds of scales
 * **Local interactions and fitness**: sparse fitness and more efficient optimization
 * **TODO and local interaction**: automatic adaptation and self-organizing clock

Multi-level modeling (methodology):
we emphasize not glueing together model at multi-scale with predefined notion of what each will do exploiting the self-organization that can take place**:**
 * BUT**
 * **simple interactions to generate complex behaviour (and its coherence)**
 * **need multi-level to understand single level phenomena:** alleviate uncertain parameters in intensive problems
 * **need to understand complex to complex mapping:** evolutionary signatures, novelty

Major transitions continued .....

 * are levels limited? in principle not, thus open-ended evolution?
 * examples:
 * autocatalytic sets versus replicating molecules where the complexity of molecules increases
 * regulatory states versus epigenetic inheritance
 * limited cultural inheritance versus language
 * here we add:
 * attractor-based versus storage-based inheritance: defined dynamics of system, but not part of dynamics
 * is this possible
 * cf interaction between gene regulation and evolution

Maynard Smith and Szathmary (1995) treat transitions at different levels as the same phenomena. We illustrate this with an example:
 * TODO and cell behaviour: gives decision making
 * TODO and grouping: in order to have coherent motion have to do what neighbour does and align
 * pick up signals from those who do something else
 * prototype is BOIDS (REF)
 * Comparing the two: persistency leads to coherences (via crowding)
 * Beltman and Maree (REF) used the CPM and persistency of cell motion
 * Cells have initial random direction
 * With persistency movement in certain direction more probable
 * Cells alone move in random direction
 * With more cells: all move in same direction and move faster
 * Thus the mechanism is very analogous to that in animal groups however:
 * there is quite a different feeling during the interpretation of the mechanisms, because for cells one needs a lot less to interpret it!

Next: Ecosystem complexity, selection and insight


 * References**
 * Maynard Smith & Szathmary** (1995) The Major Transitions in Evolution