Proposal of a modeling and testing environment for product control architectures of production systems.

Summary The development of infotronic technologies, which allow data to be carried digitally by the product in order to give it an active role in the cybernetic loop, is leading to a questioning of the conventional organization of control systems, in order to move towards a product-based control. There is a large consensus on the interest of this approach in both centralized and distributed decision making. However, few works focus on the evaluation of the efficiency of product-based control in the interaction between centralized business-level information systems (ERP) and distributed process-level systems (MES). Our contribution focuses on a modeling and simulation tool for product-controlled control systems in order to evaluate different organizational topologies combining centralized and/or distributed decisions by comparing some productivity criteria. We first present the definition, the development and the validation of a component-oriented evaluation environment, based on an emulation tool and a multi-agent system, allowing to analyze the performances of a product-controlled system and to compare it with classical approaches. We then present the application of the product-based control from a series of experiments carried out with the developed environment. These experiments, carried out on an industrial case as well as on a laboratory experimentation platform, allow us to test and validate the feasibility of the concept of product-based control in terms of decisional impact and in terms of technical constraints.