Integration in industrial automation based on multi-agent systems using cultural algorithms for optimizing the coordination mechanisms

Integration in industrial automation can be approached from the theory of Distributed Artificial Intelligence. One approach is the modeling of different production units by agents that interact through interaction protocols, which are implemented following a coordination mechanism. Under this approach, integration in automation can be achieved through the optimization of implicit interactions in such mechanisms. This paper presents a strategy for integrating industrial processes based on Multi-Agent Systems (MAS), which consists of optimizing coordination mechanisms that implement conversations between agents, by using cultural algorithms. The cultural algorithm uses formal models of interaction protocols between agents, such as auction and tender, and the integration scheme comes from automation architectures based on MAS, to which their interactions are optimized. The proposed scheme enables data and service-oriented integration. The proposed strategy is applied in two industrial case studies related to the oil production process.