Abstract
This chapter explores many aspects of the general type-2 fuzzy system that was introduced in Chap. 1. As was done for interval type-2 fuzzy systems, it provides a very comprehensive and unified description of the two major kinds of general type-2 fuzzy systems that may be used in real-world applications—GT2 Mamdani and GT2 TSK fuzzy systems. Importantly, it also distinguishes between GT2 fuzzy systems that include type-reduction followed by defuzzification and those that bypass type-reduction and use direct defuzzification.
The coverage of this chapter focuses on singleton fuzzification and the use of the horizontal-slice representation of a GT2 FS, and includes: GT2 rules, horizontal-slice formulas for firing sets and fired-rules output sets, horizontal-slice first- and second-order rule partitions, combining or not combining fired-rule output sets on the way to defuzzification, horizontal-slice type-reduction (centroid and center-of-sets) for horizontal-slice GT2 Mamdani and TSK fuzzy systems, defuzzification (this is where horizontal slices are aggregated), a summary of the computational steps for two horizontal-slice Mamdani and two horizontal-slice TSK GT2 fuzzy systems, horizontal-slice versions of the NT and BMM direct defuzzification methods, GT2 fuzzy basis functions which provide a mathematical description of a GT2 fuzzy system from its input to its output, remarks and insights about a GT2 fuzzy system, what exactly “design of a GT2 fuzzy system” means as well as a tabular way for making the choices that are needed to fully specify a GT2 fuzzy system, two approaches to design—the partially dependent approach and the totally independent approach, but only for singleton GT2 fuzzy systems—requirements that need to be met in the study of real-world applications of GT2 fuzzy systems, and a case study of GT2 fuzzy logic control. Ten examples are used to illustrate the important concepts and there is also a comprehensive numerical example in Sects. 11.9 and 11.11.
Notes
- 1.
In the Wagner and Hagras references, the term “zSlice” is used instead of horizontal-slice. See Sect. 6.7.3 for a discussion about why this book uses “horizontal-slice” instead of zSlice.
- 2.
In a GT2 fuzzy system, in (11.5) and other places, one needs to keep track of which antecedent is referred to (the first subscript i), which rule is referred to (the superscript l), and which \( \alpha{\text{-cut}} \) is referred to (the second subscript \( \alpha ) \); this unavoidably leads to very heavy subscript and superscript notations.
- 3.
Almaraashi et al. (2016) perform type-reduction followed by defuzzification by first computing the COG of vertical slices (some interpolation of the COGs is also used) and then defuzzifying the resulting T1 FS. Although this kind of type-reduction is ad hoc, it does adhere to the basic design requirement that when all MF uncertainties disappear the output of a T2 fuzzy system must become the same as the output of a T1 fuzzy system. If one is only interested in the defuzzified output of a GT2 fuzzy system, and not a measure of the uncertainties that have flowed through that system, then their vertical slice centroid type-reducer (VSCTR) is arguably a very viable alternative to the WH approach that is taken in this chapter.
- 4.
Height type-reduction for a WH GT2 fuzzy system is left as an exercise (Exercise 11.4).
- 5.
- 6.
“At least” is used here because it is conceivable that even some sort of GT2 FS functions could be used for the GT2 TSK rule consequents. How to do this is an open research issue.
- 7.
Wagner and Hagras (2010) was the first to make this observation.
- 8.
Of course, when \( \alpha \) is sampled very finely then \( y_{1} ,y_{2} ,\ldots, y_{2k} \) will approach a uniformly sampled set of samples; however, in a WH GT2 fuzzy system one does not want to use too many horizontal slices, or else computational complexity greatly increases as does computation time.
- 9.
- 10.
For notational simplification, \( y_{{l,\alpha_{k} }}^{COS} \) and \( y_{{l,\alpha_{k} }}^{COS} \) are shortened to \( y_{{l,\alpha_{k} }} \) and \( y_{{r,\alpha_{k} }} \), respectively.
- 11.
The extension of approximate type-reduction + defuzzification (Wu-Mendel uncertainty bounds), that is given in Sect. 9.8, is left to the reader as an exercise (Exercise 11.12) because its formulas are much more complicated than the ones for the extensions of the NT and BMM formulas.
- 12.
This step depends on the design choices made for the secondary MFs, and is returned to in Sect. 11.14.
- 13.
- 14.
Adaptiveness and novelty are defined in Sect. 9.13.2.
- 15.
- 16.
Recall that the FOU of a GT2 FS is its \( \alpha = 0 \) horizontal slice.
- 17.
For a different kind of design that also uses triangle secondary MFs, see Starczewski (2009). Since the advent of the horizontal-slice decomposition, it is arguably no longer necessary to use the approximate approach to type-reduction that is advocated in this paper. Instead, one only needs to design a WH GT2 fuzzy system that uses a few horizontal slices.
- 18.
- 19.
Because formulas have only been given in this chapter for singleton fuzzification, the non-singleton steps that are listed in Sect. 10.2.5 are not included here. Of course, once the formulas for T1 or IT2 non-singleton fuzzification have been obtained (Exercise 11.1) then more steps can be added to the present list, as has been done in Sect. 10.2.5.
- 20.
- 21.
- 22.
All GA designs were achieved by using the MATLABⓇ GA toolbox, and used 20 iterations and 20 populations-default MATLAB GA parameter settings. MATLAB is a registered trademark of The MathWorks, Inc.
- 23.
Note that the squishing parameters for all of these horizontal slices can be deduced from the entries into Table 11.9; each is the coefficient of the left-end of the stated interval.
References
Almaraashi, M., R. John, A. Hopgood, and S. Ahmadi. 2016. Learning of interval and general type-2 fuzzy logic systems using simulated annealing: Theory and practice. Information Sciences 360: 21–42.
Aleantara, S., R. Vilanova, and C. Pedret. 2013. PID control in terms of robustness/performance and servo/regulator tradeoffs: A unifying approach to balanced autotuning. Journal of Process Control 23 (4): 527–542.
Castillo, O., L. A.-Angulo, J. R. Castro, and M. G.-Valdez. 2016. A comparative study of type-1 fuzzy logic systems, interval tyoe-2 fuzzy logic systems and generalized type-2 fuzzy logic systems in control problems. Information Sciences 354: 257–274.
Castillo, O., L. Cervantes, J. Soria, M. Sanchez, and J.R. Castro. 2016b. A generalized type-2 fuzzy granular approach with applications to aerospace. Information Sciences 354: 165–177.
Derrac, J., S. Garcia, D. Molina, and F. Herrera. 2011. A practical tutorial on the use of nonparametric statistical tests as a methodology for comparing evolutionary and swarm intelligence algorithms. Swarm and Evolutionary Computation 1: 3–18.
Gaxiola, F., P. Melin, F. Valdez, and O. Castillo. 2015. Generalized type-2 fuzzy weight adjustment for backpropagation neural networks in time series prediction. Information Sciences 325: 159–174.
Greenfield, S., and R. John. 2009. The uncertainty associated with a type-2 fuzzy set. In Views on fuzzy sets and systems from different perspectives: Philosophy and logic, criticisms and applications, ed. R. Seising, 471–483. Heidelberg: Springer.
Hamrawi, H., and S. Coupland. 2009. Type-2 fuzzy arithmetic using alpha-planes. In Proceedings of the IFSA/EUSFLAT, 606–611. Portugal.
Karnik, N.N., J.M. Mendel, and Q. Liang. 1999. Type-2 fuzzy logic systems. IEEE Transactions on Fuzzy Systems 7: 643–658.
Kumbasar, T., and H. Hagras. 2015. A self-tuning zslices based general type-2 fuzzy PI controller. IEEE Transactions on Fuzzy Systems 23: 991–1013.
Liu, F. 2008. An efficient centroid type-reduction strategy for general type-2 fuzzy logic system. Information Sciences 178: 2224–2236.
Mendel, J.M. 2010. Type-2 fuzzy sets—A tribal parody. IEEE Computational Intelligence Magazine 5: 24–27.
Mendel, J.M. 2014. General type-2 fuzzy logic systems made simple: A tutorial. IEEE Trans on Fuzzy Systems 22: 1162–1182.
Mudi, R.K., and N.R. Pal. 1999. A robust self-tuning scheme for PI- and PD-type fuzzy controllers. IEEE Transactions on Fuzzy Systems 7 (1): 2–16.
Skogestad, S. 2006. Tuning for smooth PID control with acceptable disturbance rejection. Industrial and Engineering Chemistry Research 45 (23): 7817–7822.
Starczewski, J.T. 2009. Efficient triangular type-2 fuzzy logic systems. International Journal of Approximate Reasoning 50: 799–811.
Wagner, C., and H. Hagras. 2008. zSlices–Towards bridging the gap between interval and general type-2 fuzzy logic. In Proceedings of the IEEE FUZZ conference, Paper # FS0126. Hong Kong.
Wagner, C., and H. Hagras. 2010. Towards general type-2 fuzzy logic systems based on zSlices. IEEE Transactions on Fuzzy Systems 18: 637–660.
Wagner, C., and H. Hagras. 2013. ZSlices based general type-2 fuzzy sets and systems. In Advances in type-2 fuzzy sets and systems: Theory and applications, eds. Sadeghian, A., J. M. Mendel, and H. Tahayori. New York: Springer.
Wu, D. 2012. On the fundamental differences between interval type-2 and type-1 fuzzy logic controllers. IEEE Transactions on Fuzzy Systems 20: 832–848.
Wu, D., and W. W. Tan. 2010. Interval type-2 fuzzy PI controllers: Why they are more robust. In Proceedings of IEEE international conference on granular computing, 802–807. Silicon Valley.
Zhai, D., and J.M. Mendel. 2012. Comment on ‘toward general type-2 fuzzy logic systems based on zslices’. IEEE Transactions on Fuzzy Systems 20: 996–997.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this chapter
Cite this chapter
Mendel, J.M. (2017). General Type-2 Fuzzy Systems. In: Uncertain Rule-Based Fuzzy Systems. Springer, Cham. https://doi.org/10.1007/978-3-319-51370-6_11
Download citation
DOI: https://doi.org/10.1007/978-3-319-51370-6_11
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-51369-0
Online ISBN: 978-3-319-51370-6
eBook Packages: EngineeringEngineering (R0)