Design and conceptual development of shop floor controllers through the manipulation of process plans

Abstract

A shop-floor control system (SFCS) performs the decision-making and execution functions necessary to fill production orders efficiently. To cope with a variety of dynamic factors, the SFCS must be able to perform these functions in real time. This paper proposes a way to reduce the cost of developing the software that implements this realtime requirement. It describes a design and development framework that uses hierarchical process plans and simulation. The plans are written in the process specification language (PSL) developed at the National Institute for Standards and Technology (NIST) and coded as extensible markup language (XML) document type definitions (DTDs). The decision maker in each controller parses these DTDs into a non-linear graph, resolves the AND-junctions and OR-junctions, determines a sequence of production tasks and generates the set of messages to execute those tasks. The decision planner, when called by the decision maker, evaluates several sets of control rules using a corresponding simulation model, chooses the most promising set and provides it to the decision maker. The proposed concept enables control software to be designed and developed in terms of the evolution of process plans, so both the decision maker and the executor can be generic and the simulation model (decision planner) can be generated in an automatic manner. The generality and validity of the proposed concept has been conceptually tested for diverse manufacturing systems generated from a designed experiment.