Complexity and its control in dissociation reaction network

Structural controllability of complex systems proved [1] that a graph-theoretical technique can be used to identify the key processes through which full control of a system is achievable. This theory has been recently applied to the network of processes of a negative ion source [2,3] to test the sensitivity of the source behaviour to the driver processes. Since chemical reaction networks can be seen as graphs where reactants and products are considered as nodes of the network [4, 5](Aerts, et al., 2015), in this contribution the same technique has been applied to chemical reaction networks to identify the driver nodes of a reaction, that is, those species whose role in the reaction is particularly relevant. A network model has been generated from both a full [6] and a reduced [5] dissociation model, showing some differences between the two sets of the driver species. From a graph-theoretical point of view, these systems are thus not equivalent. To investigate this aspect, species that do not participate in both systems were removed from the network representation of the full system. The resulting network still shares more than 70% of the network structure with the network representation of the reduced model, showing that the network structure is resilient. Among the missing species, those whose removal induces differences between the driver node sets of the two models have been identified and their role discussed.