|
Author
|
Topic: Super Fast Evolution
|
warren_bergerson
Member
Member # 262
|
posted 07. May 2003 09:34
Dembski’s hypothesized role of a designer in evolutionary change is based in part on the assumption that there is maximum rate or speed of evolutionary change by natural materialistic processes. If it can be demonstrated that natural materialistic processes exist which are capable of producing one generation evolutionary changes with complexity greater than 10^150, then Dembski’s hypothesis collapses. [ It is interesting to note that while ‘super-fast’ evolution would appear contrary to Dembski’s ID hypothesis and contrary to traditional Darwinian approaches, it does not appear to conflict with current evolutionary positions. ]
Using the concept of complexity, evolutionary speed can be defined in terms of the number of generations required to find one of M survival compatible solutions in a solution space containing N possible solutions. Using Dembski’s criteria, the maximum natural speed of evolution is ‘finding a 1 in 10^150 adaptive solution in a single generation’. I suggest this arbitrary rate or speed of evolutionary change provides a useful criteria for distinguishing between ‘slow’ evolutionary change processes and ‘super-fast’ evolutionary change processes.
In an abstract mathematical universe, it is not difficult to design an super-fast evolutionary process which can evolve faster than the Dembski standard. Some of the mathematical techniques capable of generating super-fast evolutionary change would include:
1. Front loading or stored solution (the solution to the adaptive problem is already stored in the system)
2. Non Random variation- The ‘solution’ has a higher than chance probability of being considered than other members of the solution set.
3. Multiple options per generation- (search routines can simultaneously consider a whole set or range of options. As a simple example, a visual search, can consider many different locations in a single cycle. ).
4. Multiple cycles per generation-
In terms of mathematics, there is no limit to how fast evolutionary change could occur. Using any of the above techniques, or combinations of techniques, it is possible to design mathematical evolutionary processes which operate at super-fast speeds.
One of the more powerful techniques for producing super-fast evolutionary change is parallel processing. If a complex change can be separated into a set of component changes, and if the search for solutions to each of the components can progress in parallel, then speed of solving complex ‘evolutionary’ problems can be greatly increased. If parallel processing is combined with some of the other techniques list above, then super-fast evolutionary speeds could be easily achieved.
It should be apparent that it would not be difficult to build an artificial simulation of super-fast evolution using the mathematical techniques outlined. It should also be apparent that there are biological manifestations of each of the techniques listed. Natural materialistic super-fast evolution is thus at least possible.
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 08. May 2003 07:53
A couple of follow up comments. If super fast evolutionary processes exist (processes capable of solving in one generation survival problems with complexity greater than 10^150), then it would seem to follow that biological systems may have the ability or the intelligence to produce intelligent designs by natural materialistic processes.
The existence of human and animal intelligence would lend support to the argument that the physical processes needed for super fast evolution can exist in biological systems.
GA systems incorporate many of the mathematical features needed to simulate super fast evolution. It seems that GA systems should be able to simulate super fast evolution. If evolutionary capacities of GA systems are assumed to reflect evolutionary capacities of biological systems, then super fast evolution would appear to be compatible with current evolutionary theories.
On just a common sense basis, it would seem that if you wanted to construct a predictive model or theory of evolutionary change processes, the first requirement would be to define and quantify the volume and speed of evolutionary change. It would appear, at least on the surface, somewhat dubious to suggest that a predictive mathematical theory of evolution exists, if there do not even exist techniques for addressing issues like super fast evolution.
IP: Logged
|
|
|
Printer-friendly view of this topic