Structural accessibility and structural observability of networked systems, with examples from biology
Models in biology, ecology or other areas complex due to numerous co-existing species which interact, but their intractions are usually not well quantified. The state of art considers essentially linear interactions.
The analysis of those models, or rather families of models, then goes through fundamental notions as "structural" controllability or "structural" observability. In plain words, these notions characterize the controllability or observability of a generic model picked from a family of systems with linear interactions.
This talk will recall those standard results and argue the nonlinearity of the interactions. New criteria for structural accessibility and structural observability are derived for the larger families of systems which are subject to nonlinear interactions. Motivating real life examples are used to illustrate the main results.