ComputerAided_Design_Engineering_amp_Manufactur.pdf

D. Inference Engine for Expert Systems: This allows both forward and backward chaining. 11
Coding and the associated compilation are not required, thus avoiding grammatical and
typographical errors. In addition, the user can easily modify the designed inference network
in a graphical fashion. An example of the inference network for car diagnosis has been demonstrated.
Rules can be represented by transitions and preconditions, or symptoms can be
represented by tokens in places. Clicking the “Auto” button of the “Simulate” submenu of
the “Analysis” menu fires the production rules (transitions) and, when the firing terminates,
the places holding tokens indicate the cause of the problem.
8.9 Conclusions
The tool, based on the synthesis rules and the algorithm, helps designers to construct large PNs
interactively and to synthesize an automated manufacturing system in a user-friendly fashion.
None of the existing tools integrate drawing, file manipulation, analysis, simulation, animation,
reduction, synthesis, and property query in one software package. Furthermore, because PNs model
discrete-event systems, the tool finds applications in communication protocols, flexible manufacturing
systems, (extended) finite state machines, expert systems, interactive parallel debuggers, 11 digital signal
processing, 5,7,11 etc.
We have enhanced the tool to include models not only of PNs but also of state diagrams and data
flow graphs (DFGs) with few code changes. Thus a designer can choose the model with which he
is familiar. For instance, DSP professionals do not know PNs well. They can, however, draw DFGs
and obtain iteration bounds, critical loops, rate-optimal scheduling, etc. by just clicking a button. 4,5,65
We have also implemented a reduction algorithm based on the rules; the code is very simple,
containing less than 100 lines. The distinct point of this approach is that, besides the possibility of
continuous enhancement, while reducing, it can discover wrong designs and suggest how to fix the
problem based on the knitting rules.
This work overcomes some drawbacks of most existing synthesis approaches; i.e., they do not
• Deal with the algorithm and CAD tool using graphical user interface for synthesis explicitly
• Show how to continuously update their synthesis techniques
• Indicate how to extend the synthesis for analysis
• Show temporal relationships among processes after synthesis
• Find the maximum concurrency of the synthesized net.
References
1. Agerwala, T. and Y. Choed-Amphai, A synthesis rule for concurrent systems, Proc. of Design Automation
Conference, 1978, pp. 305–311.
2. Berthelot, G., Checking properties of nets using transformations, in Advances in Petri Nets, G.
Rozenberg (ed.), 1985, Springer-Verlag, pp. 19–40.
2a. Chao, D.Y., Application of a synthesis algorithm to flexible manufacturing systems, Journal of
Information Science and Engineering, Vol. 14, No. 2, June 1998, pp. 409–477.
3. Chao, D. Y. and D. T. Wang, Synchronized choice ordinary Petri net, (Invited) Proc. 1995 IEEE
Int’l Conf. SMC, Vancouver, Canada, October 22–25, pp. 1442–1447.
4. Chao, D. Y. and D. T. Wang, XPN-FMS: A modeling and simulation software for FMS using
Petri nets and X windows,” International Journal of Flexible Manufacturing Systems, Vol. 7, No.
4, October 1955, pp. 339–360.
5. Chao, D. Y. and D. T. Wang, Iteration bounds of single-rate data flow graphs for concurrent
processing, IEEE Trans. Circuits Syst., CAS-40, No. 9, September 1993, pp. 629–634.
6. Chao, D. Y. and D. T. Wang, Two theoretical and practical aspects of knitting techniques—invariants
and a new class of Petri net, IEEE Trans. SMC_27, No. 6, December 1997, pp. 962–977.