Nat. Phys. https://doi.org/10.1038/s41567-021-01474-y (2022)
Our world consists of multi-scale complex systems — such as ecological systems at the geographic scale, electric power grids at the city scale, and regulatory occurrences among genes at the molecular scale — that can be better understood by developing network models. The function of these network systems — for instance, the ecological sustainability, the energy distribution in an energy grid, and gene expression levels — depends on the self-dynamics of each node and their interactions. These network components and structures are vulnerable to topological perturbations that lead to a network dysfunction, namely, the transition from an active state (x1) to a collapsed state (x0). Even with a successful reconstruction of the topological damage that corresponds to the missing nodes (xi) or links in the network, the recovery of its function is very challenging, mainly due to the fact that the system cannot instantly reinstate its lost functionality. One conventional approach to address this issue is to apply large perturbations that dramatically drive all activity components (xi) in the network to a basin of attraction, where the system can find itself in a given stable state, such as the functional state x1. However, this all-node control can be ineffective and impractical in real systems — for instance, it would be very difficult (if not impossible) to change the status of every species (node) in an ecological network. In a recent work, Baruch Barzel and colleagues proposed an efficient recovery scheme that only requires perturbing a small set of nodes in order to drive a dysfunctional network back into a functional state.
This is a preview of subscription content, access via your institution