**SWARM INTELLIGENCE** **some questions and remarks**  QUESTION ============================================================== Are there laws linking the complexity of individuals in a (homogeneous) swarm to the maximum possible complexity of the swarm behavior (be it building something, or simply the spatial dynamics of the swarm)? NOMOLOGICAL-DEDUCTIVE EXPLANATIONS ============================================================== I'm not aware of much published research in this direction that was successful at formulating generalizable universal laws or even "explanation sketches" that Carl Hempel (https://en.wikipedia.org/wiki/Carl_Gustav_Hempel) would be happy with. Some argued that perhaps such laws could not be written down due to computational irreducibility (c.f. "A New Kind of Science" by Stephen Wolfram https://www.goodreads.com/en/book/show/238558 or "Complexity: A Guided Tour" by Melanie Mitchell https://www.goodreads.com/book/show/5597902). However, it seems such laws must exist. Human individuals are themselves swarms of less complex individuals (the definition of an individual will vary depending on the level of hierarchy we want to look at, but let's say cells). The cells face two major limitations: their own computational horizon (how much compute is available to the cell as an individual) and the topology and bandwidth of communication network with other cells (determined by biochemical processes). HINTS ============================================================== There might be some laws (perhaps yet to be formulated) in style similar to limits imposed on physical systems by relativity and quantum theory. Those could be of two kinds: communication bandwidth between parts of the swarm and perception resolution (computational limitations) of individual units within the swarm. I believe Stephen Wolfram might have already hinted at some of those in his physics project, where you can derive something similar to relativity and parts of quantum mechanics from the "ruliad", but I believe there are different ways of looking at this and some laws could be written down based on those two parameters. ANTS ============================================================== Take the ant colonies as an analogy. Depending of the compute available to the individual ants (how smart a single individual is) and their ability to communicate (speed and topology of information flow between individual nodes), I would predict there are some limits to the complexity of behaviors that the colony as a whole can generate. CIVILIZATIONS ============================================================== Another analogy is human civilizations. If we keep the compute available to an average individual constant (say we consider modern humans) and vary the information flow - by changing technology: no language -> spoken language only -> writing -> the internet -> direct brain interfaces ... we will see different levels of complexity and emergent phenomena (maybe the whole system becomes conscious at some point in this progression) at scale. EVOLUTION ============================================================== Maybe evolution in complex systems can be considered beyond the individual. Individuals are like nodes in a graph and the system evolves both on the level of nodes and on the level of connections and their weights between individual nodes. NEAT :)(https://nn.cs.utexas.edu/downloads/papers/stanley.ec02.pdf).