### Faced with the difficulty of calculating the evolution of complex systems defined by
- a great diversity of objects and
- a great diversity of interactions,
### **The gem-graph rationale is:**
#### 1. represent space
#### 2. a discreet (non-continuous) space
#### 3. a space whose all units are similar
#### 4. links can be established between some of these units
- they allow drawing **objects** (isolated connex parts of the graph) and **situations** (relative positions of objects)
- for practical purposes, it is convenient to use **arrows** and to allow stacking many of them from the same node to another
#### 5. an automaton, i.e. a set of states and transitions can rewrite this space
#### 6. ***states*** can represent ***space***
- here, **space** can be understand as **a representation or approximation of a real space**
- but a **state** can be a **space** as well as a text or any other set of symbols or writings (ex: tags) that can be drawn in the graph using the same encoding and a **state** can be an association of both.
#### 7. transitions are all combinations of ***a single type of elementary transition***
#### 8. the coding of static information (states) and dynamic information (transitions) is distinct
- the purpose of this restriction is to maintain a strict **homogeneity of the rules** (cf. §7) which is the condition of their automatic management and edition.
#### 9. no constraint on granularity: the scope of the arrows between space units is not limited
#### 10. the computation is local, random, asynchronous
#### 11. interfaces are possible with statistical and / or continuous fermion models:
- spaces are then superimposed and conditions on intensive local variables (concentrations, temperatures, flows, etc...) can be added to the specific conditions of the gem-graph.
#### 12. interfaces are possible with representations of bosons:
- spaces are then superimposed and conditions on intensive local variables (flux, cross section, etc.) can be added to the specific conditions of the gem-graph.
#### 13. the topology, the dimension and the magnitude of the space are not constrained
#### Design: two converging, complementary and interdependent approaches:
- one goes **from** the available data structures and algorithms **to** the graph structure and the automaton it supports (synthetic or bottom-up approach)
- the other goes **from** the gem-graph automaton constraints **to** the available software tools (analytical or top-down approach)