A bio-inspired emergent control approach for distributed processes

Abstract

The complexity of industrial processes has grown exponentially, with high degrees of dependency, non-linearity, imprecision, among other aspects, which motivates the interest in developing distributed control systems for their management. In this sense, this work proposes a bio-inspired distributed control approach, where control actions emerge from the component interactions. The distributed control approach is based on the response threshold model to solve the control problem by imitating the behavior of ants. Particularly, our approach is inspired by the way as the ants carry out the division of labor in a colony. Thus, our control approach based on the threshold response model refers to the possibility of reacting to stimuli associated with the distributed control tasks. It has the ability to stabilize the process in the presence of abrupt/successive changes and various initial conditions, with a minimum effort of the actuators to achieve the objectives. Also, it has shown its versatility in different operational contexts with the same parameter tuning. The bio-inspired control approach is proved in a quadruple tank process, a complex system due to its multivariate nature. In this way, our paper introduces a new domain of application of the response threshold model in industrial processes. Several experiments were carried out in different contexts to evaluate its stability, robustness, etc., and compare it with other similar works. In general, the control performance metrics show satisfactory results, which reflects its ability to adapt to changes in the dynamics of the process, which encourages additional studies.