Systemics of Complexity
Inside the galaxy of the system centred theories, the Systemics of Complexity (Le Moigne, Morin) - or SC - focuses on the qualitative study of complex systems, systems which can be of an unspecified nature (physical, biological or social). Contrary to the traditional methods known as Cartesian or analytical, its aim is not to predict (to calculate) the evolution of a complex system, but rather to make it understandable.
It is frequent to confuse complicated system and complex system. Often, this confusion is caused by the same great number of elements which compose the two types of systems and by the difficulty of understanding the interactions occurring between their elements. Moreover, in both cases, the system term means a representation or a pre-representation of the studied phenomenon. However, there is a difference in nature between these two categories of systems: a complicated system can be simplified skillfully in order to be able to carry out a particular calculation, whereas a complex system by any handling under the penalty of losing the intelligibility of the studied unit.
One of the consequences of the SC paradigm is that the systemician proposes a a possible model of the complex system and not a final mathematical function. This possible model reflects the modelisator’s project and specifies at the same time the limits of the project and of the model. Because it is a possible model, there is not a one to one relationship between a complex system and a model supposed to represent it: several complementary models can coexist in the representation (or construction) of a complex system.
We identified two types of approaches usable by an observer to study a complex system: the equational approach and the structural approach. The choice between these two procedures will be mainly guided by the objective to reach: calculate or understand. The main characteristics of these two methodologies can be stated in the following way:
- It does not take into account the internal composition of the system’s elements.
- The internal composition of an element is not regarded as relevant in the study of the complexity of the global system: the constitution of an element (internal architecture, components) is transparent for the observer.
- The elements are many and indistinguishable: they are considered as particles.
- It uses powerful mathematical formalisms.
- It favours obtaining quantitative results.
- It takes into account the internal composition of the system’s elements.
- The elements are not numerous.
- The internal composition of an element is considered as a source of complexity of the global system.
- It adapts to the presence of heterogeneous elements.
- It leans on systemic theories.
- It favours obtaining qualitative results.
Computing disciplines essentially adopt the equational approach to model complex systems. Because of the problematics studied, the ARAL team has chosen the stuctural approach. This one was used to study artificial autonomy and to define a general architecture of a property oriented software agent (Systemion Model).