From rehfuess@obsidian.zfe.siemens.de Thu Nov 10 22:11:58 EST 1994 Article: 3266 of comp.ai.fuzzy Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!pipex!uunet!Germany.EU.net!Munich.Germany.EU.net!zfe.siemens.de!obsidian.zfe.siemens.de!not-for-mail From: rehfuess@obsidian.zfe.siemens.de (Uli Rehfuess) Newsgroups: comp.ai.fuzzy Subject: SIEFUZZY - a CAE tool for fuzzy system design Date: 7 Nov 1994 14:20:43 GMT Organization: Siemens AG, The "neu-Perlach" branch / Munich-Germany-Europe. Lines: 101 Message-ID: <39ld3r$ok4@ztivax.zfe.siemens.de> NNTP-Posting-Host: obsidian.zfe.siemens.de X-Newsreader: TIN [UNIX 1.3 941006BETA PL0] Hi, fuzzy netters, recently there were quite a few questions in comp.ai.fuzzy concerning tools for the design of fuzzy systems. So I would like to draw your attention to SIEFUZZY V1.0, our CAE tool for fuzzy system design. SIEFUZZY is a software development tool which provides a way to describe fuzzy logic systems, test the systems through simulation, compile the system for further online testing, or compile the finished system into the output C-code necessary for the final implementation on the target processor. The implemented system can define the complete application or can be integrated into an existing application. SIEFUZZY can be used to develop fuzzy systems for control, information processing, expert systems, financial forecasting, decision-aiding tools, or any other area in which fuzzy logic is applicable. SIEFUZZY has been jointly developed by Togai InfraLogic, Inc., and Siemens AG. "TILShell 3.0 Professional Edition" is the Togai InfraLogic, Inc., proprietary name for SIEFUZZY. Fuzzy logic systems are rule-based systems which generally consist of a collection of rules of the following form: If (A is HIGH) AND (B is MED) THEN C = MED where A and B are the input variables to the system, C is the output variable, and HIGH and MED are fuzzy sets, or membership functions, defined on the input and output variables. SIEFUZZY provides editors for creating each part of a fuzzy logic rulebase: variables, membership functions, and rules. These editors allow you to enter descriptions of variables, membership functions, rules, and the connections between variables and rulebases in a graphical, point and click environment. The description of a fuzzy system in SIEFUZZY is maintained in the Fuzzy Programming Language (FPL TM). Each time a change is made through one of the editors in SIEFUZZY, the FPL description of the system is updated to reflect the changes made. Once the system has been completely described, it can be tested using the various debugging and tuning facilities of SIEFUZZY. These facilities include static and dynamic testing facilities. For static testing, you can enter specific values for the inputs of your system and then observe the values of the outputs, or plot a 3-dimensional representation of the relationship between two inputs and one output of your system over a given range of input values. With a cursor you can move around in the control surface and identify exactly the rules and membership functions responsible for e.g. a spike in that surface. For dynamic testing, you can test your system as it runs with a simulation of the final application and observe the behavior of each rule, membership function, and their interactions. Or as a final test, you can link SIEFUZZY via a serial connection to an external processor which is actually running the final application, and make changes to the system "online", i.e. without recompilation, and observe the results. Both in simulation and in online tuning trigger conditions can be set to detect exeptional states of the fuzzy system. The backtracking mode allows an analysis of the behaviour that led into the exeptional state of the system. SIEFUZZY includes more advanced features like DDE support, user defined extensions of FPL and user defined utilities. SIEFUZZY supports Dynamic Data Exchange (DDE) under Windows. This feature allows you to link SIEFUZZY as either a client or server to other applications, which support DDE, e.g MatLab, as well as linking two instances of SIEFUZZY. SIEFUZZY also supports user defined fuzzy operators, objects, and inference/defuzzification methods. This feature allows you to test systems which use your own specialized methods using the simulation facility. SIEFUZZY supports user defined utilities. This feature allows you to call other applications from a SIEFUZZY menu command. Prices of SIEFUZZY are: single user licence: DM 3980,- university single user licence: DM 2900,- If you need further information on SIEFUZZY, please contact our distributor at the following address: Siemens AG ANL 441 Mr. Nitsche G.-Scharowsky-Str. 2 91058 Erlangen Germany Tel: +49 9131 7 33319 Fax: +49 9131 7 26643 Mr. Nitsche will be pleased to send you detailed information and a free demo diskette of SIEFUZZY. If there are further questions via email please feel free to contact me. Sincerely U. Rehfuess _/ _/ _/_/_/ Ulrich Rehfuess _/ _/ _/ _/ Siemens AG, ZFE T SN 4 _/ _/ _/_/_/ Otto-Hahn-Ring 6 _/ _/ _/ _/ D-81730 Munich, Germany _/_/ _/ _/ e-mail: Ulrich.Rehfuess@zfe.siemens.de