Analysis and Assessment of Multi-Agent Systems for Production Planning and Control

Industrial manufacturing systems are controlled by a large number of components hierarchically organized into a so called automation pyramid. Changes to systems’ dynamic and structure are difficult to manage because of many technological interdependencies that might be broken during reconfiguration. Recent requirements for fast adaptation to customer demands and volatility of supply chains demand for new modular reconfigurable production lines. While at the moment Industry 4.0 focuses more on standardized information access and sharing, flexible control technology is still weak. Encapsulation of production functions within autonomous manufacturing modules enables new control architectures and strategies. In this work we reviewed the existing Multi-Agent Systems (MASs) for Production Planning and Control to access their applicability to the Industry 4.0 scenarios. From a big corpus of identified publications in the field of industrial MAS applications we picked only the solution-based MAS approaches that focus on production planning and control. For this paper five MAS architectures were selected for analysis and comparison according to the specified classification criteria. Though all the analyzed architectures are applicable to modular production, non of them was implemented as a standard in an actual industrial enterprise. This shows that further work on maturing the agent-based technologies in real industrial environments is needed.