subsection_Rich_models_As_all__.tex

\subsection{Rich models}

As all modelling approaches, also NGO is about simplification and abstraction. Nevertheless, the model should be formulated such that there are opportunities for a rich dynamics and hence for exploration of those dynamics. This is most likely in a bottom-up, or mechanistic approach [Hogeweg2011pcb]: one defines the building blocks and how these may interact. As an example, organisms could be the building blocks, while the rule for interaction formalises that organisms like to be close to each other and adjust their direction of movement to the direction of their close neighbours [REFS]. Next, one explores the dynamics and patterns that form as a result of these interacting components. Usually this part of the study is at a more coarse-grained spatio-temporal scale: it is not the individual that we are interested in, it is the collective behaviour of these individuals. Our experience is that often unforeseen consequences arise. In the example case, it was found that with three simple rules realistic and complex flocking patterns of birds or schools of fish can be simulated. At the time, this came as a complete surprise. Such grouping patterns were expected to involve many factors and complicated rules [REFS]. (We are not claiming that animals flock according to three simple rules. Only that flocking is an easily obtained phenomenon that may be accomplished in many ways, as few simple rules already ``do the trick''.) As Turing said in the introduction of his famous paper on morphogenesis [Turing1952], a model is a simplification and idealization, and consequently a falsification. However, a good model has interesting features that help us critically evaluate our current state of knowledge.