Situation Management of Multi-Stage Chemical-Technological Processes Using Fuzzy Co-Evolutionary Metaheuristics

The complexity of multi-stage line designs and the increasing proportion of complex technological objects in their composition have significantly complicated their management processes. Numerous external factors affecting such facilities necessitate a rapid response to changes in the current situation, and a large number of quantitative-qualitative parameters, combined with information incompleteness make it impossible to create and use analytical models for their management. This, in turn, leads to the use of situational management methods capable to determine the necessary impacts according to the current situation. The article proposes a new approach for determining the optimal control action in the current situation, using data from "situation-action" knowledge base and based on hybridization of fuzzy logic and a coevolutionary algorithm. Fuzzy logic methods enable forming fuzzy representation of the current situation, which used to determine required parametric control action based on the available precedent base. As part of dynamic optimization of control parameters, a co-evolutionary algorithm inspired by firefly interaction patterns is proposed, where solution is found through competition among multiple populations.