Macro-types in multi-scale feedback systems: a conceptual framework

Large Complex Adaptive Systems (CAS) operate at multiple scales - processing various information flows, encoded at different granularities, over different substrates. Often, system scales are interconnected by feedback cycles, with micro-scales generating macro-scales, in turn influencing the micro-scales, and so on. E.g. foraging ants (micro) contribute to a pheromone trail (macro), which in turn guides their movement (micro) and updates the trail (macro). As these processes take widely different forms when occurring in different environments and accomplishing different functions, they may be difficult to detect, compare and transfer across CAS domains. We previously generalised such diverse processes, defining them explicitly as a Multi-Scale Feedback Systems (MSFS) design pattern. In this paper, we propose a conceptual framework for categorising macroprocesses in MSFS. This helps identify and analyse micro-macro feedback cycles across CAS domains; and engineer new CAS with specific constraints. We illustrate these concepts via CAS examples that fit the MSFS pattern via different variants. This contribution extends the theoretical basis of MSFS for dealing with the increasing complexity of modern environments.