Abstract:Controllability is significant for dynamical systems, and iterative recovery of controllability is indispensable sometimes. We consider any large, sparse Erdós-Rényi random digraph with a linear time-invariant control model as an input graph, obtained by adding one node to an original random digraph, and we seek to recover its controllability via efficiently identifying a maximum matching of it rather than recomputation. Particularly, we assume that any input graph contains a known matching that is a maximum matching of the original random digraph. In our solution, we depend on a bipartite graph one-to-one mapped by an input digraph to find its augmenting paths relative to a matching corresponding to the known matching of the input graph. By finding augmenting paths within this mapped bipartite graph, we eventually find a maximum matching and recover the controllability of this input digraph in linear time as a result.

Iterative Recovery of Controllability Via Maximum Matching.