Abstract
This chapter is the IT2 version of Chap. 4. It focuses first on what exactly “design of an IT2 fuzzy system” means, and then provides a tabular way for making the choices that are needed in order to fully specify an IT2 fuzzy system. It introduces two approaches to design, the partially dependent approach and the totally independent approach, but this time for singleton, T1 non-singleton and IT2 non-singleton IT2 fuzzy systems. It then describes the extension of the six design methods that were covered for type-1 fuzzy systems in Chap. 4 to IT2 fuzzy systems, namely: IT2 WM, least squares, derivative-based, SVD-QR, derivative-free, and iterative. It continues the three Chap. 4 case studies (forecasting of time series, knowledge mining using surveys, and fuzzy logic control), and continues the Chap. 4 applications of forecasting of compressed video traffic using IT2 Mamdani and TSK fuzzy systems, and IT2 rule-based classification of video traffic. The application of equalization of time-varying nonlinear digital communication channels is also covered. Thirteen examples are used to illustrate the important concepts.
Notes
- 1.
- 2.
Mendel (2004) is all about computing derivatives for IT2 Mamdani fuzzy system with COS type-reduction + defuzzification. Because it does not use the formulation of \( y({\mathbf{x}}) \) that is given in (9.172)–(9.176), but instead uses the formulation of \( y({\mathbf{x}}) \) that is in terms of the switch points \( L({\mathbf{x}}) \) and \( R({\mathbf{x}}) \), it is very complicated. This is due to having to reorder the IT2 FBFs, something that is not needed when (9.172)–(9.176) are used.
- 3.
This is a somewhat speculative section because only the SVD-QR method that is described for the IT2 Mamdani fuzzy system with COS type-reduction + defuzzification has been successfully applied in Liang and Mendel (2000c), although Example 10.9’s Steps 4 and 5 are not explicitly mentioned in that paper, even though they were used.
- 4.
If an IT2 fuzzy system cannot be designed by means of optimization, but can only be designed by trial and error, then it is possible that the performance of the singleton IT2 fuzzy system may be worse than that of a non-singleton type-1 fuzzy system, or the performance of the T1 non-singleton IT2 fuzzy system may be worse than that of a singleton IT2 fuzzy system, or the performance of the IT2 non-singleton IT2 fuzzy system may be worse than that of a T1 non-singleton IT2 fuzzy system. This occurs because it may not be possible to try all possible combinations of the design parameters, and is therefore the result of an incomplete design procedure. One must be very cautious about drawing conclusions from such trial and error designs.
- 5.
In everything that has been described prior to this section the type-reduced set has only been a means to an end, the end being the defuzzified output. Although this author has also used the type-reduced set in this way, he always felt that it contained valuable information about dispersion about the output and that somehow this information ought to also be used in a design, since dispersion about the mean is used in probability based designs.
- 6.
Cara et al. (2013) compare a singleton IT2 fuzzy system with a non-singleton T1 fuzzy system for nine function approximation problems to address a criticism of IT2 fuzzy systems that they have more parameters (design degrees of freedom) than a T1 fuzzy system. Both fuzzy systems are constrained to have the same number of parameters. The IT2 fuzzy system uses the FOU to handle input uncertainties, whereas the T1 fuzzy system uses non-singleton fuzzification to do this. The parameters of both kinds of fuzzy system are optimized. The paper demonstrates that the IT2 fuzzy system gives much better performance than the non-singleton T1 fuzzy system for high noise levels, and that it is not the use of extra parameters that lets a singleton IT2 fuzzy system outperform a non-singleton T1 fuzzy system, but rather their different ways of handing such uncertainty.
- 7.
If the singular value threshold \( \gamma \) had been chosen to be smaller than 1, then the rule-reduced designs would have contained more rules, but would have achieved even smaller RMSEs than in the present designs.
- 8.
Note that, for \( N\widetilde{V}L \), \( m_{a} = \sigma_{a} = 0 \), so the first two coordinates of both its UMF and LMF are (0, 0) and (0, 1), as they should be; and, for \( M\widetilde{A}A \), \( m_{b} = 10 \) and \( \sigma_{b} = 0 \), so the last two coordinates of both its UMF and LMF are (10, 1) and (10, 0), as they should be.
- 9.
Note that \( \underline{C}_{avg}^{i} \) and \( \bar{C}_{avg}^{i} \) play the roles of \( c_{l} (\widetilde{G}^{i} ) \) and \( c_{r} (\widetilde{G}^{i} ) \), respectively.
- 10.
See, also, Sect. 5.2.
- 11.
Another approach for obtaining an IT2 FS word model from the same kind of data that are used by the IA, EIA and HMA is in Tahayori and Sadeghian (2012). Its Median Interval Approach (MIA) is based on calculating the median boundaries of the range of MFs associated with the words.
- 12.
A T1 FS can be used either to model each subject’s intra-uncertainty, or to model the inter-uncertainty of a group of subjects (by using group statistics), but it cannot model both kinds of uncertainties simultaneously.
- 13.
For those who are note interested in the specifics of this application, just view Fig. 4.17 as a time series, where “time” is the same as “frame index.”
- 14.
Earlier it was argued that such a TSK fuzzy system is better referred to as a Mamdani fuzzy system; however, because there is no unnormalized Mamdani fuzzy system, but there is an unnormalized TSK fuzzy system, the designation “TSK” is used here.
- 15.
- 16.
Definitions of the terms used in this section can be found in any standard textbook on communication theory, e.g. Proakis (1989).
- 17.
Wikipedia/Adaptive equalizer: accessed on June 27, 2016.
- 18.
Work that has been done in the area of adaptive equalization, but mainly for time-invariant channels, includes Proakis (1989) and the many references therein, and, Chen et al. (1993a, b, 1995), Cowan and Semnani (1998), Lee (1996), Moon and Jeon (1998), Patra and Mulgrew (1998), Sarwal and Srinath (1995), Savazzi et al. (1998), and Wang and Mendel (1993).
- 19.
See footnote 14.
- 20.
In the K-NNC, if the number of training prototypes is N, then \( K{ = }\sqrt N \) is the optimal choice for K. It is required that N be an odd integer; hence, the choice of \( N = 121 \).
- 21.
Some of the material in this section is taken from Mendel et al. (2014, Chap. 1).
- 22.
This section was prepared by Prof. Tufan Kumbasar.
- 23.
KM algorithms were used to compute the COS type-reduced set; however, the same numerical results would have been obtained had the EKM or EAISC algorithms been used.
- 24.
- 25.
Better performance may have been obtained for the three other IT2 FPID controllers if the parameters of their FOUs were tuned for each of them.
- 26.
- 27.
In these references the partitions are not referred to as IT2 first-, IT2 second- or IT2 novelty rule partitions. These names and explanations only occurred to this author during the writing of this book.
- 28.
The statement of this exercise is taken or adapted for the most part from Mendel et al. (2014, Chap. 5, pp. 207–208).
References
Bargiela, A., and W. Pedrycz. 2003. Granular computing: An introduction. Dordrecht: Kluwer Academic Publishers.
Cara, A.B., C. Wagner, H. Hagras, H. Pomares, and I. Rojas. 2013. Multiobjective optimization and comparison of nonsingleton type-1 and singleton interval type-2 fuzzy logic systems. IEEE Transcations on Fuzzy Systems 21: 459–476.
Castillo, O., and P. Melin. 2014. A review on interval type-2 fuzzy logic applications in intelligent control. Information Sciences 279: 615–631.
Chen, S., B. Mulgrew, and S. McLaughlin. 1993a. A clustering technique for digital communications channel equalization using radial basis function network. IEEE Transaction on Neural Networks 4: 570–579.
Chen, S., B. Mulgrew, and S. McLaughlin. 1993b. Adaptive Bayesian equalizer with decision feedback. IEEE Transactions on Signal Processing 41: 2918–2927.
Chen, S., S. McLaughlin, B. Mulgrew, and P.M. Grant. 1995. Adaptive Bayesian decision feedback equalizer for dispersive mobile radio channels. IEEE Transactions on Communications 43: 1937–1956.
Cowan, C.F.N., and S. Semnani. 1998. Time-variant equalization using a novel non-linear adaptive structure. International Journal of Adaptive Control and Signal Processing 12: 195–206.
Dereli, T., A. Baykasoglu, K. Altun, A. Durmusoglu, and I. Burkhan Turksen. 2011. Industrial applications of type-2 fuzzy sets and systems: A concise review. Computers in Industry 62: 125–137.
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.
Du, X., and H. Ying. 2010. Derivation and analysis of the analytical structures of the interval type-2 fuzzy-PI and PD controllers. IEEE Transactions on Fuzzy Systems 18: 802–814.
Garcia, S., A. Fernandez, J. Luengo, and F. Herrera. 2010. Advanced nonparametric tests for multiple comparisons in the design of experiments in computational intelligence and data mining: Experimental analysis of power. Information Sciences 180: 2044–2064.
Hagras, H. 2007. Type-2 FLCs: A new generation of fuzzy controllers. IEEE Computational Intelligence Magazine 2: 30–43.
Hagras, H., and C. Wagner. 2012. Towards the wide spread use of type-2 fuzzy logic systems in real world applications. IEEE Computational Intelligence Magazine 7 (3): 14–24.
Hao, M., and J.M. Mendel. 2015. Encoding words into normal interval type-2 fuzzy sets: HM method, accepted for publication in IEEE Transactions on Fuzzy Systems.
Karnik, N.N., J.M. Mendel, and Q. Liang. 1999. Type-2 fuzzy logic systems. IEEE Transactions on Fuzzy Systems 7: 643–658.
Lee, K.Y. 1996. Complex fuzzy adaptive filters with lms algorithm. IEEE Transactions on Signal Processing 44: 424–429.
Liang, Q. and J.M. Mendel. 2000a. Decision feedback equalizer for nonlinear time-varying channels using type-2 fuzzy adaptive filters,” In Proceedings of FUZZ-IEEE’00, San Antonio, TX, May 2000a.
Liang, Q., and J.M. Mendel. 2000b. Equalization of nonlinear time-varying channels using Type-2 fuzzy adaptive filters. IEEE Transactions on Fuzzy Systems 8: 551–563.
Liang, Q., and J.M. Mendel. 2000c. Designing interval type-2 fuzzy logic systems using an SVD–QR method: Rule reduction. Int’ernational Journal of Intelligent Systems 15: 939–957.
Liang, Q. and J.M. Mendel. 2001. MPEG VBR video traffic modeling and classification using fuzzy techniques. IEEE Transactions on Fuzzy Systems, pp. 183–193, Feb. 2001.
Liu, F., and J.M. Mendel. 2008. Encoding words into interval type-2 fuzzy sets using an Interval Approach. IEEE Transactions on Fuzzy Systems 16: 1503–1521.
Magdon-Ismail, M., A. Nicholson, and Y. Abu-Mostafa. 1998. Financial markets: Very noisy information processing. Proceedings of IEEE 86: 2184–2195.
Marín, L.G., F. Valencia and D. Sáez. 2016. Prediction interval based on type-2 fuzzy systems for wind power generation and loads in microgrid control design. In Proceedings FUZZ-IEEE 2016, Vancouver, Canada, July 2016.
Mendel, J.M. 2000. Uncertainty, fuzzy logic, and signal processing. Signal Processing Journal 80: 913–933.
Mendel, J.M. 2001. Introduction to rule-based fuzzy logic systems. Upper Saddle River, NJ: Prentice-Hall.
Mendel, J.M. 2003. Fuzzy sets for words: A new beginning. In Proceedings of 2003 IEEE int’ernational conference on fuzzy systems, St. Louis, MO, pp. 37–42, May 2003.
Mendel, J.M. 2004. Computing derivatives in interval type-2 fuzzy logic systems. IEEE Transactions on Fuzzy Systems 12: 84–98.
Mendel, J.M. 2007a. Advances in type-2 fuzzy sets and systems. Information Sciences 177: 84–110.
Mendel, J.M. 2007b. Computing with words and its relationships with fuzzistics. Information Sciences 177: 998–1006.
Mendel, J.M. 2007c. Computing with words: Zadeh, Turing, Popper and Occam. IEEE Computational Intelligence Magazine 2: 10–17.
Mendel, J.M., H. Hagras, W.-W. Tan, W.W. Melek, and H. Ying. 2014. Introduction to type-2 fuzzy logic control. Hoboken, NJ: John Wiley and IEEE Press.
Mendel, J.M., and R.I. John. 2002. Type-2 fuzzy sets made simple. IEEE Transactions on Fuzzy Systems 10: 117–127.
Mendel, J.M., and D. Wu. 2010. Perceptual computing: Aiding people in making subjective judgments. Hoboken, NJ: Wiley and IEEE Press.
Mendel, J.M. 2014. General type-2 fuzzy logic systems made simple: A tutorial. IEEE Transactions on Fuzzy Systems 22: 1162–1182.
Moon, J., and T. Jeon. 1998. Sequence detection for binary ISI channels using signal space partitioning. IEEE Transactions on Communications 46: 891–901.
Nie M., and W.W. Tan. 2010. Derivation of the analytical structure of symmetrical IT2 fuzzy PD and PI controllers. In Proceedings of 2010 IEEE int’ernationl conference on Fuzzy Systems, Barcelona, Spain, pp. 1–8, July 2010.
Patra, S.K., and B. Mulgrew. 1998. Efficient architecture for Bayesian equalization using fuzzy filters. IEEE Transactions Circuits and Systems II: Analog and Digital Signal Processing 45: 812–820.
Proakis, J.G. 1989. Digital communications, 2nd ed. New York: McGraw-Hill.
Rickard, J.T., J. Aisbett, R. Yager and G. Gibbon. 2011. Linguistic weighted power means: Comparison with the linguistic weighted average. In Proceedings FUZZ-IEEE 2011, 2011 World Congress on Computational Intelligence, pp. 2185–2192, Taipei, Taiwan.
Rickard, J.T., J. Aisbett, R.R. Yager, and G. Gibbon. 2013. Computing with words using weighted power mean aggregation operators. In Soft computing: State of the art theory and novel applications, ed. R.R. Yager, A.M. Abbasov, M.Z. Reformat, and S. Shahbazova, 145–160. New York: Springer.
Rutkowski, L. 2004. Flexible neuro-fuzzy systems: Structures, learning and performance evaluation. Boston: Kluwer.
Sarwal, P., and M.D. Srinath. 1995. A fuzzy logic system for channel equalization. IEEE Transactions Fuzzy Systems 3: 246–249.
Savazzi, P., L. Favalli, E. Costamagna, and A. Mecocci. 1998. A suboptimal approach to channel equalization based on the nearest neighbor rule. IEEE Journal Selected Areas in Communications 16: 1640–1648.
Tahayori, H., and A. Sadeghian. 2012. Median interval approach to model words with interval type-2 fuzzy sets. International Journal of Advanced Intelligence Paradigms 4 (3): 313–336.
Wang, L.-X., and J.M. Mendel. 1993. Fuzzy adaptive filters, with application to nonlinear channel equalization. IEEE Transactions Fuzzy Systems 1 (3): 161–170.
Wu, D., J.M. Mendel, and S. Coupland. 2012. Enhanced interval approach for encoding words into interval type-2 fuzzy sets and its convergence analysis. IEEE Transactions on Fuzzy Systems 20: 499–513.
Wu, D., and W.W. Tan. 2006. Genetic learning and performance evaluation of type-2 fuzzy logic controllers. Engineering Applications of Artificial Intelligence 19 (8): 829–841.
Wu, H., and J.M. Mendel. 2007. Classification of battlefield ground vehicles using acoustic features and fuzzy logic rule-based classifiers. IEEE Transactions on Fuzzy Systems 15: 56–72.
Zhou, H., and H. Ying. 2013. A method for deriving the analytical structure of a broad class of typical interval type-2 Mamdani fuzzy controllers. IEEE Transactions on Fuzzy Systems 21: 447–491.
Zhou, S.-M., J.M. Garibaldi, R.I. John, and F. Chiclana. 2009. On constructing parsimonious type-2 fuzzy logic systems via influential rule selection. IEEE Transactions on Fuzzy Systems 17: 654–667.
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). Interval Type-2 Fuzzy Systems: Design Methods and Applications. In: Uncertain Rule-Based Fuzzy Systems. Springer, Cham. https://doi.org/10.1007/978-3-319-51370-6_10
Download citation
DOI: https://doi.org/10.1007/978-3-319-51370-6_10
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-51369-0
Online ISBN: 978-3-319-51370-6
eBook Packages: EngineeringEngineering (R0)