KSME Journal, Vol. 9, No.3, pp. 269-285, 1995 Performance Management of Integrated Manufacturing System Networks Sulk Lee*, Asok Ray**, Keum Shik Hong*, Joongsun Yoon* and Myung Chul Han* (Received April 28, 1994) Performance management of communication networks is critical for speed, reliability, and flexibility of information exchange between different components, subsystems, and sectors (e.g., factory, engineering design, and administration) of production process organizations in the environment of Computer Integrated Manufacturing (CIM). The objective is to improve the efficiency in handling various types of messages, e.g., control signals, sensor data and production orders, by on-line adjustment of the parameters of the network protocol. This paper presents conceptual design, development, and implementation of a performance management procedure for CIM applications. The performance management algorithm is formulated using the concepts of: (i) perturbation analysis of discrete event dynamic systems; (ii) stochastic approximation; and (iii) learning automata. The performance management procedure has been tested via emulation on a network testbed that is based on the Manufacturing Automation Protocol (MAP). The conceptual design presented in this paper offers a step forward to bridging the gap between management standards and users' demands for efficient network operations since most standards such as ISO and IEEE address only the architecture, services, and interfaces for network management. The proposed concept for performance management can also be used as a general framework to assist design, operation, and management of flexible manufacturing systems. Key Words: Computer Networks, Performance Management, Computer Integrated Manufac- turing 269 1. Introduction Availability of affordable computer hardware and software has resulted in automation of vari- ous of1fice and factory operations. These opera- tions include accounting, forecasting, and market- ing for business administration; Computer Aided Design (CAD), Finite Element Analysis (FEA), and computer simulations for design and engi- neering; and, Computer Aided Process Planning (CAPP), inventory control, and monitoring & control of production processes for manufactur- *Research Institute of Mechanical Technology, Pusan National University, Pusan, Korea **Department of Mechanical Engineering, The Pen- nsylvania State University, University Park, PA, U.SA. ing. In an effort for further productivity gain, Computer Integrated Manufacturing (CIM) focuses on combining these activities into a single, closed-loop, and interactive control system. Essential to the success of CIM is computer networking which links spatially distributed islands of automation through timely exchange of relevant information such as new production orders, design modifications, and status report of manufacturing resources (Ray, 1988). Computer networks for CIM are required to transfer information under continuous changes such as fluctuations in network trafIic including addition, deletion, and failure of network compo- nents. In order to maintain an acceptable level of service, computer networks must be flexible to adapt themselves to variations in traffic load; this is the responsibility of network management. In