Application of the immune system network concept to sequential control

Demonstrates the application of the immune system reaction network concept to sequential control, by determining control sequences of sequential control plants that can be modeled by a sub-class of Petri nets with modular modeling characteristics, defined as PI*-net. The determination of control sequences of a sequential control plant modeled by Petri nets is equivalent to the determination of firing sequences of these models. We propose an approach for finding firing sequences based on the immune system response mechanism, which differs substantially from conventional enumerative methods. The firing sequence problem is reformulated by decomposing it into several smaller sub-problems, the model itself is divided into several sub-models and local firing sequences that will compose the original model firing sequence are selected by interacting common transitions of these sub-models, through local objective functions. Local objective functions are set dynamically by internal and inter sub-problems interactions, based on the immune system response regulatory model. Simulations demonstrated the validity and promising performance of the proposed approach.