Meta-Models for Integrating Production Management Functions in Heterogeneous Industrial Systems

Several heterogenous processes can be found in industrial practice, such as plants involving partly continuous, partly discrete manufacturing and companies composed by departments managed by different production planning philosophies. An efficient management of heterogeneity calls for decentralization. This paper addresses a distributed approach to job-shop scheduling in discrete manufacturing, by analysing conditions for a robust integration of a set of autonomous multiple agents for the shop-floor control in a time-varying manufacturing environment. Instead of machines processing jobs according to a plan established by a global controller, a population of intelligent local controllers should cooperatively operate in order to achieve both individual goals (the best possible utilization of machining and transportation resources) and the overall system target (minimum waiting time).