【正文】 is to enhance our fundamental understanding of design theory and methodology by extending it to the field of engineering systems design. A system design theory for largescale and plex system is not yet fully developed. Particularly, the question of how to simplify a plicated or plex design task has not been tackled in a scientific way. Furthermore, building a bridge between design theory and the latest epistemological outes of formal representations in puter sciences and operations research, such as discrete event system modeling, can advance future development in engineering design. Application in Logical Hardware Design From a logical perspective, PLC software design is similar to the hardware design of integrated circuits. Modern VLSI designs are extremely plex with several million parts and a product development time of 3 years [Whitney, 1996]. The design process is normally separated into a ponent design and a system design stage. At ponent design stage, single functions are designed and verified. At system design stage, ponents are aggregated and the whole system behavior and functionality is tested through simulation. In general, a plete verification is impossible. Hence, a systematic approach as exemplified for the PLC program design may impact the logical hardware design. PLC PLC（ programmable logical controller） face ever more plex challenge these days. Where once they quietly relays and gave an occasional report to a corporate mainframe, they are now grounded into cells, give new jobs and new languages, and are forced to pete against a growing array of control products. For this year’ s annual PLC technology update, we queried PLC makers on these topics and more. Programming languages Higher level PLC programming languages have been around for some time, but lately their popularity has been mushrooming. As Raymond Lavelle, vice president and general manager, Siemens Energy and Automation. Inc, Programmable Controls Division, points out :” As programmable controls are being used for more and more sophisticated operations, languages other than ladder logic bee more practical, efficient, and powerful. For example, it39。 IEC 61499, 1998。1 PLC 英文資料 Motivation Programmable Logic Controllers (PLC), a puting device invented by Richard E. Morley in 1968, have been widely used in industry including manufacturing systems, transportation systems, chemical process facilities, and many others. At that time, the PLC replaced the hardwired logic with softwired logic or socalled relay ladder logic (RLL), a programming language visually resembling the hardwired logic, and reduced thereby the configuration time from 6 months down to 6 days [Moody and Morley, 1999]. Although PC based control has started to e into place, PLC based control will remain the technique to which the majority of industrial applications will adhere due to its higher performance, lower price, and superior reliability in harsh environments. Moreover, according to a study on the PLC market of Frost and Sullivan [1995], an increase of the annual sales volume to 15 million PLCs per year with the hardware value of more than 8 billion US dollars has been predicted, though the prices of puting hardware is steadily dropping. The inventor of the PLC, Richard E Morley, fairly considers the PLC market as a 5billion industry at the present time. Though PLCs are widely used in industrial practice, the programming of PLC based control systems is still very much relying on trialanderror. Alike software engineering, PLC software design is facing the software dilemma or crisis in a similar way. Morley himself emphasized this aspect most forcefully by indicating [Moody and Morley, 1999, p. 110]: `If houses were built like software projects, a single woodpecker could destroy civilization.” Particularly, practical problems in PLC programming are to eliminate software bugs and to reduce the maintenance costs of old ladder logic programs. Though the hardware costs of PLCs are dropping continuously, reducing the scan time of the ladder logic is still an issue in industry so that lowcost PLCs can be used. In general, the productivity in generating PLC is far behind pared to other domains, for instance, VLSI design, where efficient puter aided design tools are in practice. Existent software engineering methodologies are not necessarily applicable to the PLC based software design because PLCprogramming requires a simultaneous consideration of hardware and software. The software design bees, thereby, more and more the major cost driver. In many industrial design projects, more than SO0/a of the manpower allocated for the control system design and installation is scheduled for testing and debugging PLC programs [Rockwell, 1999]. In addition, current PLC based control systems are not properly designed to support the growing demand for flexibility and reconfigurability of manufacturing systems. A further problem, impelling the need for a systematic design methodology, is the increasing software plexity in largescale projects. Objective and Significance of the Thesis The objective of this thesis is to develop a systematic software design methodology for PLC operated automation systems. The design methodology involves highlevel description based on state transition models that treat automation control systems as discrete event systems, a stepwise design process, and set of design rules providing guidance and measurements to achieve a successful design. The tangible oute of this research is to find a way to reduce the uncertainty in managing the control software development process, that is, reducing programming and debugging time and their variation, increasing flexibility of the automation systems, and enabling software reusability through modularity. The goal is