Structure discrimination of continuous models for (bio)chemical reaction networks via finite state machines

In chemical engineering and biology, reaction networks play a paramount role because of the complexity of such systems. In this paper, we suggest a procedure that allows discarding certain reaction schemes without going through the identification step. We use a finite state machine, which we generate directly from the reaction scheme. Its behaviour covers the behaviour of the corresponding set of differential equations (for any possible value of the parameter vector). Thus, if the automaton fails to explain the experimental data, so will the ODE systems and the reaction scheme under consideration can be discarded.