Int J Adv Manuf Technol (1992) 7:168--177 (~ 1992 Springer-Verlag London Limited Intm'nltional JoumM of I]dvanced manufacturing Technologg A Review of Petri-Net Applications in Manufacturing J. A. Cecil, K. Srihari and C. R. Emerson Department of Mechanical and Industrial Engineering, T. J. Watson School of Engineering, Applied Science and Technology, State University of New York, Binghamton, NY 13902-6000, USA Petri-nets (PNs) can model concurrent and synchronous activities in a manufacturing system at various levels of abstraction. They have been used for modelling manufacturing systems, knowledge representation on the shop floor, process-plarming applications, decision- support tasks, etc. PNs are being used in a growing variety of application areas. This paper focuses on PN applications in manufacturing. A comprehensive review of research is provided. The paper also describes the design and application of PNs in the modelling of a manufacturing cell the representation of the working of the cell at various levels' of" abstraction, and the inferences that can be drawn through PN use. Ideas for future research are presented. Keywords: Petri nets; Manufacturing modelling; Cell control; Petri net applications 1. Introduction Petri-nets (PNs) are being increasingly applied in the modelling, analysis and control of discrete manufactur- ing systems [1-5]. The ability of PNs to model concurrent, synchronised interactions within a manu- facturing system has contributed to their development as a powerful modelling tool. PNs describe a manufac- turing system graphically, and this contributes to a better understanding of the complex interactions within the system [1]. Petri-net theory deals with systems-organisation theory, which had its roots in C. A. Petri's dissertation in 1962. Since then, PNs have been modified, extensions proposed and detailed theoretical analysis performed. PNs are a "model for procedures, organisation and devices where regulated flows play a role" [6]. In PN Accepted for publication: 8 January 1991 Correspondence and offprint requests to: K. Srihari, Ph.D, Assistant Professor, Department of Mechanical and Industrial Engineering, State University of New York, Binghamton, New York 13902-6000, USA. modelling, independent events and causal dependencies are represented explicitly by introducing concurrent relations, which form the conceptual basis of net theory [6]. Systems can be represented at various levels of abstraction using the same descriptive language. PN representations allow for system-property verification in a systematic manner. This paper presents an overview of basic PN theory, and discusses the use of PNs in a manufacturing scenario. It describes the modelling capabilities of PNs in manufacturing and highlights their application in knowledge representation (KR). The role of PNs in decision-support activities as well as in the analysis of collision-avoidance strategies for material-handling robots in an automated work cell are covered. The various extensions to PNs are reviewed. A detailed example of the modelling of an automated manufactur- ing facility with multiple machining centres is provided. The advantages and limitations of using PNs to model systems are elaborated, and future research areas involving Pns are addressed. 2. Components of a Petri-Net Model A PN consists of places (P) and transitions (T), which are linked to each other by arcs (A). PNs can be described as bipartite directed graphs whose nodes are a set of places and a set of transitions. PN components are presented in Fig. 1. Places. Graphically, places are represented by circles (round nodes - see Fig. 1). Places represent passive system components (resources), which store "items" (called tokens) and take particular states [6-7]. Places are also referred to as "S-elements" (from the German for place, Steller). Transitions. Rectangular boxes or bars represent transitions, which are the active system components. These components may produce, transport and change "items" [6, 7]. Transitions are also referred to as "T- elements". For each transition, there is a set of input