A Multi-agent-system for Decision Support with Petri-Nets in Large Production Systems

Very complex production systems with millions of requirements and changes at very short intervals cannot be controlled in real time in an economically optimal manner. Existing computer systems cannot adequately model the problem and cannot calculate a solution in real time. To solve this problem a special scenario-based model has been developed at the Ostfalia University of Applied Sciences, in Salzgitter–Germany, to control all simultaneous requirements in a big production scenario time in an economically acceptable manner. For instance, the delivering of material, the transportation of goods, the warehousing of material and the movement of the forklifts can be described in the same model. This model is the foundation for a decentralized software-system, which can be realized with smartphone-apps or with every other decentralized computer technology. Thus, in a complex production scenario, every chief executive could have a smartphone-app which controls his work in a decentralized manner. The software model is based on Petri-nets, which can visualize and simulate complex production scenarios. Every element in the production scenario was realized as a software agent in a multi-agent-software-system. The aim of the project is to use this software to control very complex requirements in production in real time. A live performance of a software-prototype can show the feasibility of the project.