Division of Research December 1983 Graduate School of Business Administration The University of Michigan CLASSIFICATION OF FLEXIBLE IM.NUFACTURING SYSTEMS: EVOLUTION TOWARDS THE &UTOMATED F&CTORY Working Paper No. 363 Kathryn Stecke, Didier Dubois, Jim Browne, Suresh P. Sethi and Keith Rathmill The University of Michigan FOR DISCUSSION PURPOSES ONLY None of this material is to be quoted or reproduced without the expressed permission of the Division of Research.

CLASSIFICATION OF FLEXIBLE MANUFACTURING SYSTEMS: EVOLUTION TOWARDS THE AUTOMATED FACTORY Kathryn E. Stecke Graduate School of Business Administration The University of Michigan Ann Arbor Michigan USA Didier Dubois Department d'Etudes et de Recherches en Automatique Centre d'Etudes et de Recherches de Toulouse 31055 Toulouse Cedex FRANCE Jim Browne Department of Industrial Engineering University College Galway REPUBLIC OF IRELAND Suresh P. Sethi Faculty of Management Studies University of Toronto Toronto Ontario CANADA Keith Rathmill CRAG Cranfield Institute of Technology Bedford UNITED KINGDOM December 1983 ABSTRACT There has been some uncertainty concerning whether a manufacturing system can be termed "flexible." To clarify this confusion, first, eight types of flexibilities are defined and described. Different FMSs contain different amounts of these flexibilities. We found it useful to classify types of FMSs with respect to their flexibility. The types define a range of flexibility. We observe- that an FMS's type is largely defined by its operating characteristics, but also by the versatility of its machine tools and the kind of automated material handling system. Then, an overview of the evolution of an automated factory, as an integrated combination of FMSs of different types, is presented. At each stage during the evolution, various components of automation would be implemented modularly as desired. Finally, the (future) interfaces between some of these components are discussed.

1. INTRODUCTION A flexible manufacturing system (FMS) is an integrated, computer-controlled complex of automated material handling deyices and numerically controlled (NC) machine tools that can simultaneously process medium-sized volumes of a variety of part types (Stecke [1983]). This new production technology has been designed to attain the efficiency of well-balanced, machine-paced transfer lines, while utilizing the flexibility that job shops have to simultaneously machine multiple part types. Recently, many new manufacturing facilities are labeled as FMS. This has caused some confusion about what constitutes an FMS. Flexibility and automation are the key conceptual requirements. However, it is the extent of automation and the diversity of the parts that are important; some systems are termed FMS just because they contain automated material handling. For example, dedicated, fixed, transfer lines or systems containing only automated storage and retrieval are not FMSs. Other systems only contain several [unintegrated] NC or CNC machines. Still other systems use a computer to control the machines, but often have long set-ups required or no automated part transfer. Others are called flexible because they produce a variety of parts (of very similar type, using fixed automation). In most of these examples, the operating mode is either transfer line-like or based on producing batches of similar part types. To help clarify the situation, we first define and describe eight types of flexibility in ~2. All FMSs have the potential to utilize these flexibilities to a great extent. However, there are some economic, technological, practical, and strategic reasons not to design and/or utilize all of these flexibilities in a particular FMS. For example, some FMSs operate as fixed-route transfer lines, in part because the development of real-time scheduling capabilities are difficult, expensive, and time-consuming. The level of automction helps to determine the amount of\ available flexibility. Because of the different choices of various flexibility levels, there are different types of FMSs. It is, therefore, useful to classify these systems in terms of their overall flexibility. We propose such a classification scheme for FMSs in ~3. Next, in ~4 we describe stages, in terms of an increasing amount of automation, during the development, evolution, and eventual implementation of a flexible, automated factory. Several of the components, in particular, group technology, computer aided design, and production planning and control are briefly described. Finally, the interfaces, both existing and future, between various automation components are discussed in ~5. 2. TYPES OF FLEXIBILITY Flexibility may be considered to be the most important, yet unquantifiable, aspect of an FMS. However, a considerable degree of ambiguity surrounds this term. In fact, many "FMSs" have only limited flexibility. In this section, alternative meanings of flexibility are defined. Examples or explanations are provided when needed to illustrate a particular type. Measurement and attainability of each are also discussed.

-2 - 1. Machine Flexibility: the ease of making the changes required to produce a given set of part types. Measurement of these changes include, for example, the time to replace worn-out or broken cutting tools, the time to change tools in a tool magazine to produce a different subset of the given part types, and the time to assemble or mount the new fixtures required. The setup time required for a machine tool to switch from one part type to another includes: cutting tool preparation time; part positioning and releasing time; and NC program changeover time. This flexibility can be attained by: a) technological progress, such as sophisticated tool-loading and part-loading devices; b) proper operation assignment, so that there is no need to change the cutting tools that are in the tool magazines, or they are changed less often; c) having the technological capability of bringing both the part and required cutting tools to the machine tool together. 2. Process Flexiblity: the ability to produce a given set of part types, each possibly using different materials, in several ways. Buzacott [1982] calls this "job flexibility," which "relates to the mix of jobs which the system can process." Process flexibility increases as machine set-up costs decrease. Each part can be machined individually, and not necessarily in batches. This flexibility can be measured by the number of part types that can simultaneously be processed without using batches. This flexibility can be attained by having: a) machine flexibility; and b) multi-purpose, adaptable, CNC machining centers. 3. Product Flexibility: the ability to changeover to produce a new (set of) product(s) very economically and quickly. Mandelbaum [1978] calls this "action flexibility, the capacity for taking new action to meet new circumstances." This flexibility heightens a company's potential responsiveness to competitive and/or market changes. Product flexibility can be measured by the time required to switch from one part mix to another, of not necessarily of the same part types. This flexibility can be attained by having: a) an efficient and automated production planning and control system containing: i) automatic operation assignment procedures; and ii) automatic pallet distribution calculation capability. b) machine flexibility. 4. Routing Flexibility: the ability to handle breakdowns and to continue producing the given set of part types. This ability exists if either a part type can be processed via several routes, or, equivalently, each operation can be performed on more than one machine. Note that this flexibility can be: Potential: part routes are fixed, but parts are automatically rerouted when a breakdown occurs;

-3 - Actual: identical parts are actually processed through different routes, independent of breakdown situations. The main, applicable circumstance occurs when a system component, such as a machine tool, breaks down. This flexibility can be measured by the robustness of the FMS when breakdowns occur: the production rate does not decrease dramatically and parts continue to be processed. This flexibility can be attained by allowing for automated and automatic rerouting of parts (potential routing flexibility), by pooling machines into machine groups (see Stecke and Solberg [1982]), which also allows machine tool redundancy; and also by duplicating operation assignments (see Stecke [1983]). These latter policies provide actual routing flexibility. The FMS would then be state-driven by a feedback control policy. 5. Volume Flexibility: the ability to operate an FMS profitably at different production volumes. A higher level of automation increases this flexibility, partly as a result of both lower machine set-up costs and lower variable costs. If it is not economical to run a particular system at its usual volume, say during a decrease in market demand or a recession, then there are less personnel problems concerning the idling of labor. Perhaps alternative uses of the FMS could be found. Also, production volumes can vary from week to week, resulting in variable machine and system utilizations. This flexiblity can be measured by how small the volumes can be for all part types. The lower the volume is, the more volume-flexible the system must be. This flexibility can be attained by having: a) multipurpose machines; and b) a layout that is not dedicated to a particular process; and c) a sophisticated, automated materials handling system, such as (possibly intelligent) carts, and not fixed-route conveyors; and d) routing flexibility. 6. Expandability Flexibility: the capability of building a system, and expanding it as needed, easily and modularly. This is not possible with most assembly and transfer lines. This flexibility can be measured according to how large the FMS can become. This flexibility is attained by having: a) a non-dedicated, non-process-driven layout; and b) a flexible materials handling system consisting of, say, wireguided carts; and c) modular, flexible machining cells with pallet changers; and d) routing flexibility. 7. Operation Flexibility: the ability to interchange the ordering of several operations for each part type. There is usually some required partial precedence structure for a particular part type. However, for some operations, their respective ordering is arbitrary. Some process planner has usually determined a fixed ordering of all operations, each on a

-4 - particular machine (type). However, keeping the routing options open and not pre-determining either the "next" operation or the "next" machine increases the flexibility to make these decisions in real-time. These decisions should depend on the current system state (which machine tools are currently idle, busy, or bottleneck). 8. Production Flexibility: the range of part types that the FMS can produce. This flexiblity is measured by the level of existing technology. It is attained by increasing the level of technology and the versatility of the machine tools. The capabilities of all of the previous flexibilities are required. Not all of these flexiblity types are independent. Figure 1 displays the relationships between the different flexibilities. The arrows signify "necessary for." An ideal FMS would possess all of the defined flexibilities. However, the cost of the latest in hardware and the most sophisticated (and at present nonexistent!) software to plan and control adequately would be quite high on some of these measures and low on others. For instance, processing a particular group of products may be made possible through the use of head indexers having multiple-spindle heads. However, they hinder both adding new part types to the mix and introducing new part numbers, since retooling costs are high and changeover time can be a day. Also, some flexible systems (such as the SCAMP system in Colchester, U.K.) include special-purpose, non-CNC machines, such as hobbing and broaching, which also require (relatively) huge setup times. We next use this classification of flexibilities to help categorize different types of FMSs. Product Flexibility Machine Flexibility -> Process Flexibility Operation Flexibility Production Vol~r IF^ lFlexibility Routing Flexibility olume Flexibility {Eg xpandability Flexibility FIGURE 1 Relationships Among Flexibility Types. 3. CLASSIFICATION OF FLEXIBLE MANUFACTURING SYSTEMS Towards a classification of flexible manufacturing systems, Groover [1980] divides FMSs into two distinct types: i) Dedicated FMS; ii) Random FMS. A dedicated system machines a fixed set of part types with well-defined manufacturing requirements over a known time horizon. The 'random FMS', on the other hand, machines a greater variety of parts in random sequence.

-5 - In addition to these basic, extreme types of FMSs, we note that all FMSs are different in terms of the amounts of the flexibilities that they utilize. In this section, we provide a classification of FMSs according to their inherent, overall flexibility. We will define four general types of FMS. The following standards are provided based on FMS components, which we shall use to describe and classify the different types of FMSs: 1. Machine tools: * General-purpose or specialized * Automatic tool changing capabilities (increase flexibility) * Regarding tool magazines, their capacity, removability, and toolchanging needs (affect the flexibility) 2. Materials handling system: Types include: conveyor or one-way carousel; tow-line with carts; network of wire-guided carts; stand alone robot carts Part movement equipment: palletized and/or fixtured Tool transportation system: manual; or, automatically, with parts 3. Storage areas for in-process inventory: * Central buffer storage Decentralized buffer at each machine tool * Local storage 4. Computer control: * Distribution of decisions Architecture of the information system Types of decisions: input sequence; priority rules; part to cart assignment; cart traffic regulation Control of part mix: through periodic input; through a feedbackbased priority rule. These "flexibility" standards for the physical FMS components are used to clarify differences and similarities between the FMS types. Although not typically considered FMS, our classification scheme will include the flexible assembly system (FAS). We begin by defining the simplest components, or modules, that can be considered during the construction of an FMS. The simplest possible component of an FMS or FAS is a flexible assembly cell (FAC). It consists of one or more robots and peripheral equipment, such as an input/output buffer and automated material handling. To date, only about 6% of robot application is in assembly. A flexible assembly system (FAS) consists of two or more FACs. In the future, as the technology to allow the interface between manufacturing and assembly is further developed, an FAS could also be a component of a flexible system. The types of FMS are now described. They are categorized according to the extent of use of their flexibilities. The classification of a particular FMS usually results basically from its mode of operation as well as the properties of the four components described above.

Type I FMS: Flexible MachiningC1ell The simplest, hence most flexible (especially with respect to five of the flexibilities), type of FMS is a flexible machining cell (FMC). It consists of one general-purpose CNC machine tool, interfaced with automated material handling which provides raw castings or semi-finished parts from an input buffer for machining, loads and unloads the machine tool, and transports the finished workpiece to an output buffer for eventual removal to its next destination. An articulated arm, robot, or pallet changer is sometimes used to load and unload. Storage includes the raw castings area, the input and output buffers of the machine tools, and the finished parts area. Since an FMC contains only one metal-cutting machine tool, one might question it being called a system. However, it has all of the components of an FMS. Also, it is actually an FMS component itself. With one machine tool, it is the smallest, most trivial FMS. Type II FMS: Flexible Machining System The second type of FMS can have the following features. It can have realtime, on-line control of part production. It should allow several routes for parts, with small volume production of each, and consists of FMCs of different types of general-purpose, metal-removing machine tools. Real-time control capabilities can automatically allow muttiple routes for parts, which complicate scheduling software. Because of real-time control, however, the actual scheduling might be easier. For example, the scheduling rule might be to route randomly, or route to the nearest free machine tool of the correct machine type. The scheduling rule could be some appropriate, system-dependent, dynamic-with-feedback, priority rule. Sometimes, dedicated, special-purpose machines tools, such as multiplespindle head changers, are used in an FMS to increase production. The machine tools are unordered in a process-independent layout. It is the part types that are to be processed by an FMS which define the necessary, required machine tools. A Type II FMS is highly machine-flexible, process-flexible, and productflexible. It is also highly routing-flexible, since it can easily and automatically cope with machine tool or other breakdowns if machines are grouped or operation assignments are duplicated. Within the Type II category, the various kinds of material handling provide a subrange of flexibility. In order of increasing flexibility, various material handling systems include: power roller conveyors, overhead conveyors, shuttle conveyors, in-floor tow line conveyors, and wire-guided carts. Some examples include: i) a network of carts and decentralized storage areas, for shorter processing times (Renault Machines Outils, in Boutheon, France); ii) a tow line with carts and centralized storage areas, for longer processing times (Sundstrand/Caterpillar DNC Line, in Peoria, Illinois, U.S.A.).

-7 - Type III FMS: Flexible Transfer Line The third type of FMS has the following features. For all part types, each operation is assigned to, and performed on, only one machine. This results in a fixed route for each part through the system. The layout is process-driven and hence ordered. The material handling system is usually a carousel or conveyor. The storage area is local, usually between each machine. In addition to general-purpose machines, it can contain special-purpose machines, robots, and some dedicated equipment. Scheduling, to balance machine workloads, is easier. In fact, a Type III FMS is easier to manage because it operates similarly to a dedicated transfer line. The computer control is more simple and a periodic input of parts is realistic. Once set up, it is easy to run and to be efficient. The difference is that it is set up often and relatively quickly. A Type III FMS is less process-flexible and less capable of automatically handling breakdowns. However, the system can adapt by retooling and manually inputting the appropriate command to the computer, to reroute parts to the capable machine tool. This takes more time than the automatic rerouting available to a Type II FMS. Type IV FMS: Flexible Transfer Multi-Line The fourth FMS type consists of duplicate Type III FMSs. This duplication does not increase process flexibility. Similar to a Type III FMS, scheduling and control are relatively easy, once the system is set up. The main advantage is the redundancy that it provides in a breakdown situation, to increase its routing-flexibility. It tries to achieve the best of both FMS Types II and III. Flexibility Range All things being equal, a Type II FMS is operated "flexibly," while a Type III FMS is operated in a much more "fixed" manner. These types provide the extremes, say, the bounds on flexibility. There is, of course, a whole range of flexibilities between the two general types. However, these smaller variations in flexibility are defined by the versatilities and capabilities of the machine tools, which are dictated by the particular FMS application, i.e., the part types to be machined. The types of material handling system also provides subgroups of flexiblity. The overall flexibility, however, is defined by an FMS's mode of operation. In general, the FMSs of the United States and the Federal Republic of Germany tend to be more like the Type II FMS, while those of Japan are more similar to Type III. The second floor of Fanuc's Fuji complex, consisting of four assembly lines, is an example of an operating Type IV FMS. It consists of several identical FACs, which are not all identically tooled. Parts do have fixed routes, but if an assembly cell is down, the parts requiring it are automatically able to be routed to another assembly cell, which contains the correct tooling. The first floor of this Fanuc plant, the Motor Manufacturing Division, is a good example of Type II.

-8 - All FMSs consist of similar components. The numbers and types of machine tools may differ. What really defines the flexibility of an installation is how it is run. The level of desired flexibility is an important decision in the development and implementation of an FMS. 4. EVOLUTION OF AN AUTOMATED FACTORY One general approach to modularly develop, plan, construct, and eventually implement an automated factory (AF) is provided in Figure 2. Beginning with an NC machine tool, for example, when the automated functions that are listed in Figure 2 are included (such as an automatic load and unload under CNC control capability), an FMC is the result. Then, integrating several FMCs, each possibly containing a different type of CNC machine tool, with some of the additional automated features in Figure 2, provides an FMS. The present state of the technology does not allow the automatic integration of an FMS and an FAS. Future systems should. The eventual connections and interfaces between several FMSs, FASs, with a group technology (GT) system and a computer aided design (CAD) system can provide a totally automated factory. Short descriptions of some of these other components of an automated factory, such as CAD, GT, and production planning and control are in order. Computer Aided Design Computer aided design (CAD) is both product and process related. First, the geometry of a particular part type is captured at a computer terminal and input into an engineering data base. Analytical software is used to analyze the part type design. Then, computer aided process planning (CAPP) is used to define the operations of the part type. Finally, NC part programming provides the software instructions to run the CNC machine tools that manufacture the part type. The source instructions of the NC part program provide the interface (via a post processor that is unique to the machine) between the part type geometry and its desired material, and the CNC machine tool that cuts the part. Group Technology Group technology (GT) is another design tool that is also both product and process related. First, all part types are classified and coded according to their characteristics, such as size, volume, weight, shape, material, cost, and machine tool required. This information is used to group similar part types together. This data can also be used to standardize processes and materials for similar part types and hence reduce the number of part types in the data base. Next, a cell of machine tools is chosen that is dedicated to produce that family of part types. GT is mostly applicable to a product-driven industry, rather than a process-driven. During the design of an FMS, occasionally GT has been used to help identify the appropriate families of part types to produce. This initial choice impacts the number and types of machine tools in the FMS. As part type requirements change and new part types are designed to be machined by the FMS, additional and appropriate machine tools can then be added to the FMS.

-9 - INC Machine Tool\ I Robot I I — Automatic Load and Unload* Tool Life Monitoring Pre- and Post-Storage Buffer Automated Material Transfer Tool Length Compensation* Probe (Tool and Work Gauging)* CNC Control Flexible Machining Cell __ (FMS) Flexible Assembly Cell (FAC) Automatic Palletizing and Fixturing Adaptive Control Automated Material Handling Auxiliary Equipment (Inspection Machines, Washing Station, Chip Removal,...) Automated Production Planning and Control DNC (Or Other System Control Computer Hierarchy) Flexible Manufacturing System _(FMS) Flexible Assembly System...... (FAS) General Management System Automatic Material Management Integration/Coordination of Several FMSs and FACs Automatic Storage and Retrieval of Parts, Subassemblies, Fixtures, Cutting Tools,... Integrated Design/Manufacturing Data Base Group Technology (GT) w w~~~~~~~~~~ Computer Aided Design (CAD) Automated Factory (AF) FIGURE 2 Modular Evolution to an Automated Factory. *These features apply mainly to NC machine tools, rather than to robots.

-10 - Automated Production Planning and Control Some properties and constraints of FMSs are similar to those of flow and job shops, while others are different. This technology creates the need to develop new and appropriate planning and control procedures that take advantage of the system's capabilities for higher production rates (Stecke [1983]). A complete discussion of FMS planning and control functions is beyond the scope of this paper. However, since there has been much recent research concerning planning and control models for FMSs, a bibliography of this research is provided. Flexible with Conventional Systems As an older machine tool wears out, there has been a tendency to replace it with NC. Of course, conventional manufacturing systems or lines within a plant will not disappear. Future research must also address the problem of interfacing automated with conventional systems. In addition, there will always be situations that one does not want to automate. 5. INTERFACES BETWEEN VARIOUS COMPONENTS OF AUTOMATION Some of the interfaces between the various modular components and requirements given in Figure 2 exist, while others do not yet. Recall that the NC part program is an interface between CAD and FMS. However, the automatic generation of part programs is not yet widely available. Another interface is the automatic linkage between an MRP-type output —a pick list, say —and the computer that controls the automatic storage and retrieval system or the automated warehouse. There is no interface yet between GT and MRP-type output. Robotics is still a stand alone application via the FAC and FAS. Robots cannot yet be linked to a central engineering data base from the part type geometry provided by a CAD system. All of manufacturing involves transmitting, sorting, analyzing, and modifying data. Parts are a manifestation of data. Data is a useful corporate resource. There are different design/engineering CAD data bases. Current research is investigating the problems concerning how to communicate between them. The state-of-the-art integrates CAD's geometric engineering data with the alphanumeric manufacturing information that is already in a business data base. ACKNOWLEDGEMENTS Kathryn E. Stecke's research was supported in part by a summer research grant from the Graduate School of Business Administration at The University of Michigan as well as by a grant from the Ford Motor Company, Dearborn, Michigan.

-11 - REFERENCES 1. BUZACOTT, JOHN A., "The Fundamental Principles of Flexibility in Manufacturing Systems," Proceedings of the 1st International Conference on Flexible Manufacturing Systems, Brighton, U.K. (October 20-22, 1982). 2. GROOVER, MIKELL P., Automation, Production Systems, and Computer-Aided Manufacturing, Prentice-Hall, Englewood Cliffs, NJ (1980). 2. MANDELBAUM, MARVIN, "Flexibility in Decision-Making: An Exploration and Unification," Ph.D. dissertation, Department of Industrial Engineering, University of Toronto, Ontario, Canada (1978). 3. STECKE, KATHRYN E., "Formulation and Solution of Nonlinear Integer Production Planning Problems for Flexible Manufacturing Systems," Management Science, Vol. 29, No. 3, pp. 273-288 (March 1983). 4. STECKE, KATHRYN E. and SOLBERG, JAMES, J., "The Optimality of Unbalanced Workloads and Machine Group Sizes for Flexible Manufacturing Systems," Working Paper No. 290, Division of Research, Graduate School of Business Administration, The University of Michigan, Ann Arbor MI (January 1982). 5. ZELENOVIC, D. M., "Flexibility —a Condition for Effective Production Systems," International Journal of Production Research, Vol 20, No. 3, pp. 319-337 (May-June 1982). BIBLIOGRAPHY OF FMS PRODUCTION PLANNING AND CONTROL PROBLEMS AND MODELS BARASH, MOSHE M., "Computerized Manufacturing Systems for Discrete Products," Ch. VII-9 in The Handbook of Industrial Engineering, Salvendy, G. (Editor), John Wiley and Sons, Inc., NY (1982). BERRADA, MOHAMMED and STECKE, KATHRYN E., "A Branch and Bound Approach for Machine Loading in Flexible Manufacturing Systems," Working Paper No. 329, Division of Research, Graduate School of Business Administration, The University of Michigan, Ann Arbor MI (April 1983). BUZACOTT, JOHN A., "'Optimal' Operating Rules for Automated Manufacturing Systems," IEEE Transactions on Automatic Control, Vol. AC-27, No. 1, pp. 80-86 (February 1982). BUZACOTT, JOHN A. and SHANTHIKUMAR, J. GEORGE, "Models for Understanding Flexible Manufacturing Systems," AIIE Transactions, Vol. 12, No. 4, pp. 339-349 (December 1980). BUZACOTT, JOHN A. and YAO, DAVID D. W., "Flexible Manufacturing Systems: A Review of Models," Working Paper No. 7, Department of Industrial Engineering, University of Toronto, Toronto, Ontario Canada (March, 1982).

-12 - CAVAILLE, JEAN-BERNARD and DUBOIS, DIDIER, "Heuristic Methods Based on MeanValue Analysis for Flexible Manufacturing Systems Performance Evaluation," Proceedings of the IEEE Conference on Decision and Control, Orlando FL, pp. 1061-1065 (December 1982). DUBOIS, DIDIER, "A Mathematical Model of a Flexible Manufacturing System with Limited In-Process Inventory," European Journal of Operational Research Vol. 14, No. 1, pp. 66-78 (1983). DUBOIS, DIDIER, SETHI, SURESH P. and STECKE, KATHRYN E., "A Petri Net Formulation for the Control of a Flexible Manufacturing System," presented at the Joint National ORSA/TIMS Meeting, Orlando FL (November 7-9, 1983). DUBOIS, DIDIER AND STECKE, KATHRYN E., "Using Petri Nets to Represent Production Processes," Proceedings of the IEEE Conference on Decision and Control, San Antonio TX (December 14-16, 1983). HILDEBRANDT, RICHARD R., "Scheduling and Control of Flexible Machining Systems When Machines are Prone to Failure," Ph.D. dissertation, Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge MA (August 1980). HILDEBRANDT, RICHARD R., "Scheduling Flexible Machining Systems Using Mean Value Analysis," Proceedings of the IEEE Conference on Decision and Control, Albuquerque NM (1980). HO, YU CHI and CAO, XIREN, "Peturbation Analysis and Optimization of Queueing Networks," Journal of Optimization Theory and Applications (1983). KIMEMIA, JOSEPH and GERSHWIN, STANLEY B., "Multicommodity Network Flow Optimization in Flexible Manufacturing Systems," Report No. ELS-FR-834-2, M.I.T., Cambridge MA (1978). LEVEQUE, DIDIER, ROUBELLAT, F. AND ERSHLER, J., "Periodic Loading of Flexible Manufacturing Systems," I IFIP Congress "APMS 1982," Bourdeaux, France pp. 327-339 (1982). MORTAZAVIAN, HASSAN, SETHI, SURESH P., WONHAM, MURRAY AND RAMADGE, PETER, "An Approach to the Supervisory Control of Some Manufacturing Systems," presented at the Joint National ORSA/TIMS Meeting, Orlando FL (November 7-9, 1983). NOF, SHIMON, BARASH, MOSHIE M., and SOLBERG, JAMES J., "Operational Control of Items Flow in Versatile Manufacturing Systems," International Journal of Production Research, Vol. 17, pp. 479-489 (1979). SOLBERG, JAMES J., "A Mathematical Model of Computerized Manufacturing Systems," in Proceedings of the 4th International Conference on Production Research, Tokyo, Japan (August 1977).

-13 - SOLBERG, JAMES J., "Analytical Performance Evaluation of Flexible Manufacturing Systems," Proceedings of the 18th IEEE Conference on Decision and Control, San Diego CA, pp. 640-644 (December 1979). STECKE, KATHRYN E., "Experimental Investigation of a Computerized Manufacturing System," Master's Thesis, Purdue University, W. Lafayette IN (December 1977). STECKE, KATHRYN E., "Production Planning Problems for Flexible Manufacturing Systems," Ph.D. dissertation, Purdue University, W. Lafayette IN (August 1981). STECKE, KATHRYN E., "A Hierarchical Approach to Production Planning in Flexible Manufacturing Systems," in the Proceedings of Twentieth Annual Allerton Conference on Communication, Control, and Conrputing, Monticello IL (October 6-8, 1982). STECKE, KATHRYN E. and MORIN, THOMAS L., "Optimality of Balanced Workloads in Flexible Manufacturing Systems," Working Paper No. 289, Division of Research, Graduate School of Business Administration, The University of Michigan, Ann Arbor MI (January 1982). STECKE, KATHRYN E. and SCHMEISER, BRUCE W., "Equivalent Representations of System Throughput in Closed Queueing Network Models of Multiserver Queues," Working Paper No. 324, Division of Research, Graduate School of Business Administration, The University of Michigan, Ann Arbor MI (December 1982). STECKE, KATHRYN E. and SOLBERG, JAMES J., "The CMS Loading Problem," Report No. 20, NSF Grant No. APR 74 15256, School of Industrial Engineering, Purdue University, W. Lafayette IN (February 1981a). STECKE, KATHRYN E. and SOLBERG, JAMES J., "Loading and Control Policies for a Flexible Manufacturing System," International Journal of Production Research, Vol. 19, No. 5, pp. 481-490 (September-October 1981b). SURI, RAJAN, "New Techniques for Modelling and Control of Flexible Automated Manufacturing Systems," in Proceedings, IFAC 8th Triennial World Congress, Kyoto, Japan, pp. 175-181 (1981). SURI, RAJAN AND CAO, XIREN, "Optimization of Flexible Manufacturing Systems Using New Techniques in Discrete Event Systems," Proceedings of the 20th Allerton Conference on Communication, Control and Computing, Monticello IL (October 1982).