next up previous contents index
Next: Distributed Intelligence - Related Up: ooop-thesis Previous: The Object-Oriented Ontogenetic Programming   Contents   Index


The Goal: Evolution of Distributed Intelligence

An object-oriented ontogenetic program is made of several units that interact for creating a joint behaviour that the single units are not capable of producing. No matter how we define intelligence, the object-oriented ontogenetic program can be called distributed intelligence if we have the opinion that the behaviour of the program shows intelligence. This is obvious if OOOP is used in the explicitly distributed problem solving approach described in section 2.1.3, in which the units represent and are later realized as separate physical objects that interact in the real world with messages on the basis of radio, sound, light or the like. But also in the centralized OOOP approach, in which it is used as modularization technique for a single program, the intelligence it includes can be seen as distributed. In the world of the program, the different modules are separate units that have to work together to solve a given problem. So in this case the intelligence is distributed between them.

In this context, I understand intelligence (in contrast to its common meaning in Artificial Intelligence as we will see in section 4.3) as follows:

Intelligence is a property of a system that shows intelligent behaviour which in turn cannot be intersubjectively and statically defined. If the behaviour of a system (for arbitrary reasons) is seen by a person as being intelligent, this makes the system include intelligence.
There are many programs that exhibit a behaviour that makes some people think: "Wow, that's intelligent!" With the above definition, the program will therefore have intelligence. It's creation will be a development of intelligence. But that does not mean that it will keep the honor of this characterization forever. Computer programs that have been seen as being intelligent in the 60s are now being laughed about. Animals can suffer a similar fate1. A long time, bees were marvelled at for their great cooperation and building abilities. But the more we find simple behavioural rules for single bees that can create such a cooperation, the less we are tempted to call their behaviour intelligent.

As we are evolving programs that solve problems which we ourselves cannot solve and in a way that we cannot understand, we will be tempted to call their behaviour intelligent and can therefore call the evolution of such programs an evolution of intelligence (or of distributed intelligence in the case of OOOPS).



Subsections
next up previous contents index
Next: Distributed Intelligence - Related Up: ooop-thesis Previous: The Object-Oriented Ontogenetic Programming   Contents   Index
 
© 2002 Peter Schmutter (http://www.schmutter.de)