This paper surveys an application of fuzzy set theory to process control. This application is considered as one example of rule based decision making, hence the emphasis is on an implementation of a general approach to rule based decision making. A fuzzy logic controller is described, in which the control policy is derived from control rules. These fuzzy control rules are linguistic conditional statements relating the input variables of controller to output variables. These rules are obtained from an experienced operator and reflect his intuitive feel for the process. An example of cement kiln is discussed and the control rules of fuzzy logic controller are compared with a protocol for controlling cement kiln in other studies.
The initial set of rules for fuzzy logic controller is often not the best set of rules and hence there is a need to “tune” the controller. The tuning can be done either by specifying new rules and deleting old rules, or by changing the values of scaling parameters which map real values onto a universe of discourse. Only a small change on performance can be achieved by the scaling parameters and a knowledge of process and instrumentation can generally provide adequate guide to parameter selection. Thus, the most effective way of modifying controller is to change the control rules. This can be done if the performance of the process is given in terms of reinforcements required to achieve satisfactory performance. A Self-Organizing Controller, having these features is described.
In many applications of rule based decision making the final decision (which in fuzzy logic is a fuzzy subset on output universe of discourse) is often desirable in linguistic form. This would then form the basis of a conversational program. The problem of linguistic approximation, that is giving a linguistic label to output subset, is stated. The linguistic label attached to any particular subset may consist of one or more primary phrases and these primary phrases may also be modified by hedges. A brief methodology used in constructing a final linguistic phrase is discussed. The usefulness of linguistic approximation is also considered.
KeywordsFuzzy Logic Fuzzy Logic Controller Fuzzy Subset Decision Table Control Rule
Unable to display preview. Download preview PDF.
- Buchanan, B. G., and Lederberg, J., 1971, The Heuristic DENDRIL Program for Explaining Empirical Data, Proc. of the IFIP-1971 Congress.Google Scholar
- Eshragh, F. and Mamdani, E. H., 1978, A General Approach to Linguistic Approximation, Internal Report, Dept. of Elec. Eng., Queen Mary College, London. (Also to appear in Int. J. Man-Machine Studies, 1979).Google Scholar
- Peray, K. E. and Waddell, J. J., 1972, “The Rotary Cement Kiln,” The Chemical Publishing Company, New York.Google Scholar
- Procyk, T. J., and Mamdani, E. H., 1979, A Linguistic Self-Organising Process Controller, Automatica (January).Google Scholar
- Shortliffe, E. H., 1976, “Computer-Based Medical Consultations: MYCIN,” Elsevier.Google Scholar
- Sembi, B. S. and Mamdani, E. H., 1979, On the Nature of Implication in Fuzzy Logic, Proc. 9th Int. Symp. on MVL, Bath 1979.Google Scholar
- Wenstoep, F., 1976, Deductive Verbal Models of Organisations, Int. J. Man-Machine Studies, 8.Google Scholar
- Zadeh, L. A., 1973, Outline of a New Approach to the Analysis of Complex Systems and Decision Processes, IEEE Trans. SMC-3, p. 28.Google Scholar
- Zadeh, L. A., 1975, Calculus of Fuzzy Restrictions, in: “Proc. US Open Seminar on Fuzzy Sets and their Applications,” L. A. Zadeh, K. S. Fu, K. Tanaka, M. Shimura, eds., Academic Press, New York.Google Scholar