Emergence, Locating and Measuring It - Systematic Design for Emergence in Cellular Nonlinear Networks - page 46

4

Emergence, Locating and Measuring It

4.1 Emergence: The Software Engineering of Cellular
Computing Systems

Cellular computing systems hold the potential of modeling and understanding

of complex aspects of our world. So far the CNN paradigm [7], which integrates the

cellular automata (CA) as a special case, offers a unified framework to understand

complex emergent behaviors and to apply them in real world applications.
The issue of emergence [15] in the form of a relationship between the structure

of the cell and the global behavior was identified as a key factor in understanding

cellular systems. To date emergence is vaguely defined and the most recent defini-
tions perceive it as a “surprise effect” [54], i.e. a global behavior which cannot be
predicted by simply knowing the local structure of the cell. Among emergent be-
haviors, there are some which may be effectively exploited for various applica-
tions. Since our final goal is to use cellular automata for such applications, a first
step is to precisely quantify emergent behaviors and give hints for using the meas-
ures of emergence to select (filter) only those CAs with meaningful computational
properties. This chapter and the next two focuses on these issues.
The immediate conclusion of the “surprise effect” definition for emergence is
that one in principle cannot know whether there is something interesting in the
dynamics of a cellular system without running a simulation of the system and ob-
serving the global behavior. Though, recent results based on the “local activity
theory” [8] suggest that, at least for reaction-diffusion cellular systems, knowing
the structure of the local cell may give a good prediction for the likelihood of
emergent behaviors [15], without running the entire system. Still, the exact nature
of the behavior and its relevance for computational applications has to be estab-
lished only after effectively running the cellular system and observing its behav-
ior. Therefore it is reasonable to investigate what are the relationship between
emergence and the local structure of the cell, at least for reducing the search space
through all possible cells to a much manageable subset. But the first step in start-
ing to do such analysis is to quantify emergence. This is the subject of this and the
next two chapters.
As mentioned previously, emergence is defined usually based on observation.
The main problem is how is this observation done ? For instance, the method of