computational intelligence software for interval type-2 fuzzy logic
TRANSCRIPT
Computational IntelligenceSoftware for Interval Type-2Fuzzy LogicOSCAR CASTILLO,1 PATRICIA MELIN,1 JUAN R. CASTRO2
1Tijuana Institute of Technology, Calzada Tecnologico S/N, 22379 Tijuana, Mexico
2UABC University, Calzada Universidad S/N, 22379 Tijuana, Mexico
Received 15 February 2010; accepted 21 December 2010
ABSTRACT: A software tool for interval type-2 fuzzy logic is presented in this article. The software tool includes
a graphical user interface for construction, edition, and observation of the fuzzy systems. The Interval Type-2 Fuzzy
Logic System Toolbox (IT2FLS) has a user-friendly environment for interval type-2 fuzzy logic inference system
development. Tools that cover the different phases of the fuzzy system design process, from the initial description
phase, to thefinal implementationphase, arepresentedaspart of theToolbox. TheToolbox’s best properties are the
capacity to develop complex systems and the flexibility that permits the user to extend the availability of functions
for working with type-2 fuzzy operators, linguistic variables, interval type-2 membership functions, defuzzification
methods, and the evaluation of interval type-2 fuzzy inference systems. The toolbox can be used for educational and
research purposes. �2011 Wiley Periodicals, Inc. Comput Appl Eng Educ; View this article online at wileyonlineli-
brary.com; DOI 10.1002/cae.20522
Keywords: type-2 fuzzy logic; fuzzy systems; Fuzzy Inference System Design
INTRODUCTION
Fuzzy sets were originally presented by Zadeh in 1965 [1,2] to
process/manipulate data and information affected by un-probabil-
istic uncertainty/imprecision. These were designed to mathemat-
ically represent the vagueness and uncertainty of linguistic
problems, thereby obtaining formal tools to work with intrinsic
imprecision in different types of problems [3].
Intelligent systems based on fuzzy logic are fundamental
tools for nonlinear complex system modeling. Fuzzy sets and
fuzzy logic are the basis for fuzzy systems, where their goal
has been to model how the brain manipulates inexact information.
Type-2 fuzzy sets are used for modeling uncertainty and impre-
cision in a better way. These type-2 fuzzy sets were originally
presented by Zadeh in 1975 and are essentially ‘‘fuzzy fuzzy’’ sets
where the fuzzy degree of membership is a type-1 fuzzy set [4–6].
The new concepts were introduced by Karnik and Mendel [7] and
Liang and Mendel [8] allowing the characterization of a type-2
fuzzy set with a superior membership function and an inferior
membership function; these two functions can be represented each
by a type-1 fuzzy set membership function. The interval between
these two functions represents the footprint of uncertainty (FOU),
which is used to characterize a type-2 fuzzy set. Recently, there
have been a lot of applications of type-2 fuzzy systems and for this
reason we considered the need of building a software tool that
could ease the development of type-2 fuzzy systems for real-world
problems [9,10]. Also, there is an imperative need of having this
software tool for teaching graduate and undergraduate students the
basic concepts and methods of type-2 fuzzy logic. For these
reasons, we consider that the interval type-2 fuzzy logic toolbox
makes an important contribution both to education and research of
type-2 fuzzy logic.
Currently, most of the tools for designing fuzzy systems
are focused on conventional type-1 fuzzy logic [9]. However,
for type-2 fuzzy logic there are only a few attempts of providing
tools that can help in the design of type-2 fuzzy systems. The main
reasons for not having widely used tools in this area are that type-2
fuzzy logic has only been used for a few years and also that type-2
fuzzy systems aremore difficult to design due to the higher number
of parameters involved [9,10]. As a consequence, our goal was to
design and develop a friendly and efficient type-2 fuzzy logic
toolbox that could be used both for research and education pur-
poses. The contribution of the research work has been to facilitate
the use of the theory and concepts of type-2 fuzzy logic for both
researchers and students. Regarding researchers, there have been
several type-2 fuzzy systems applications that have been achieved
by using the toolbox. Also, on the educational side there has been a
significant improvement in the level of knowledge of the students
that have used the toolbox, which is appreciated with a performed
evaluation to a sample of students.
Correspondence to O. Castillo ([email protected]).
� 2011 Wiley Periodicals, Inc.
1
INTERVAL TYPE-2 FUZZY SET THEORY
In this section the basic concepts of interval type-2 fuzzy logic
are presented. Also, the main differences with respect to conven-
tional type-1 fuzzy logic are explained. This section is very
important to understand the need of having the software presented
in this article.
Type-2 Fuzzy Sets Concept
A type-2 fuzzy set [7,8] expresses the non-deterministic truth
degree with imprecision and uncertainty for an element that
belongs to a set. A type-2 fuzzy set, denoted by~~A, is charac-
terized by a type-2 membership function m~~Aðx; uÞ, where
x2X;u2Jux � ½0; 1� and 0 � m~~Aðx; uÞ � 1 aredefined inEquation(1).
~~A ¼ ðx;m~~AðxÞÞjx2X
n o
~~A ¼ ðx; u;m~~Aðx; uÞÞj8x2X; 8u2Jux � ½0; 1�
n o(1)
An example of a type-2 membership function constructed
with the IT2FLS toolbox is composed of a Pi primary and a Gbell
secondary type-1 membership functions, and is depicted in
Figure 1.
If~~A is continuous then it is denoted as in Equation (2).
~~A ¼Zx2X
Zu2Jux�½0;1�
fxðuÞ=u
264
375=x
8><>:
9>=>; (2)
whereR R
denotes the union of x and u. If~~A is discrete then it is
denoted by Equation (3).
~~A ¼Xx2X
m~~AðxÞ=x
( )¼
XNi¼1
XMi
k¼1
fxiðuikÞ=uik" #
=xi
( )(3)
wherePP
denotes the union of x and u. If fxðuÞ ¼ 1;
8u2½Jux ; Jux � � ½0; 1�, the type-2 membership function
m~~Aðx; uÞis expressed by one type-1 inferior membership function,
Jux � mAðxÞ and one type-1 superior, Jux � mAðxÞ (Fig. 2), then it is
called an interval type-2 fuzzy set [8] denoted by Equations (4)
and (5).
~~A ¼ ðx; u; 1Þj8x2X;8u2½m
AðxÞ;mAðxÞ� � ½0; 1�
�(4)
or
~~A ¼Zx2X
Zu2½Jux ;J
u
x ��½0;1�
1=u
2664
3775=x
8>><>>:
9>>=>>;
¼Zx2X
Zu2½m
AðxÞ;mAðxÞ��½0;1�
1=u
264
375=x
8><>:
9>=>; (5)
If~~A is a type-2 fuzzy Singleton, then the membership func-
tion is defined by Equation (6):
m~~AðxÞ ¼ 1=1 si x ¼ x0
1=0 si x 6¼ x0
�(6)
Figure 1 FOU for type-2 membership functions. [Color figure can be viewed in the online issue, which is available at
wileyonlinelibrary.com.]
2 CASTILLO, MELIN, AND CASTRO
Interval type-2 fuzzy sets are the ones that, at themoment, are
mostwidely used in the applications due to their simplicity and less
computer overhead requirements. Generalized type-2 fuzzy sets
are, as of today, too complex to perform the computations and as
consequence are not widely used [9]. Figure 2 shows the FOU for
several types of interval type-2 membership functions.
Fuzzy Set Operations
We can apply some operators to the fuzzy sets, or we can make
some operations between them [4,10,11]. When we apply an
operator to one fuzzy set we obtain another fuzzy set; by the same
manner when we combine an operation with two or more sets we
obtain another fuzzy set. If we have two type-2 fuzzy subsets
identified by the letters~~A and ~~B, associated with a linguistic
variable, then we can define three basic operations: complement,
union and intersection (Table 1).
Fuzzy Inference System
The human knowledge can be expressed as fuzzy rules with the
following form [12]:
IF <fuzzy proposition> THEN <fuzzy proposition>
Figure 2 FOU for Gbell primary interval type-2 membership functions. [Color figure can be viewed in the online issue,
which is available at wileyonlinelibrary.com.]
Table 1 Interval Type-2 Fuzzy Set Operations
Name Operator Operation
UnionR ¼ join ~~A~~B ¼ R
x2Xm~~A
ðxÞm~~BðxÞ
� �¼ R
x2X
Ra2½m~A
ðxÞ_m~BðxÞ;m~AðxÞ_m~BðxÞ�
1=a
24
35=x
8<:
9=;
Intersection ¼ meet ¼ Rx2X
R�2½�
~AðxÞ^�
~BðxÞ;�~AðxÞ^�~BðxÞ�
1=�
24
35=x
8<:
9=;
Negation : :~~A ¼ Rx2X
m~~AðxÞ=x
� �¼ R
x2X
Ra2½1�m~A
ðxÞ;1�m~AðxÞ�
1=a
24
35=x
8<:
9=;
COMPUTATIONAL INTELLIGENCE SOFTWARE 3
The fuzzy propositions are divided into two types, the first
one is called atomic: x is A, where x is a linguistic variable andA is
a linguistic value; the second one is called composite: x is A AND
y is BORz is NOTC, this is a composite atomic fuzzy proposition
with the ‘‘AND,’’ ‘‘OR,’’ and ‘‘NOT’’ connectors, representing
fuzzy intersection, union, and complement, respectively. The
composite fuzzy propositions are fuzzy relationships. The mem-
bership function of the rule IF-THEN is a fuzzy relation deter-
mined by a fuzzy implication operator. The fuzzy rules combine
one or more fuzzy sets of entry, called antecedent, and are associ-
ated with one output fuzzy set, called consequents. The fuzzy sets
of the antecedent are associated with fuzzy operators AND, OR,
NOT, and linguistic modifiers. The fuzzy rules permit expressing
the available knowledge about the relationship between antecedent
and consequents. To express this knowledge completely we nor-
mally have several rules, grouped to form what it is known a rule
base, that is, a set of rules that express the known relationships
between antecedent and consequents. The fuzzy rules are basically
IF <Antecedent> THEN <Consequent> and express a fuzzy
relationship or proposition.
In fuzzy logic the reasoning is imprecise (approximate),
which means that we can infer from one rule a conclusion even
if the antecedent does not comply completely.We can count on two
basic inference methods between rules and inference laws, Gener-
alized Modus Ponens (GMP) [5,6,8,13] and Generalized Modus
Tollens (GMT), that represent the extensions or generalizations of
classic reasoning. The GMP inference method is known as direct
reasoning and is summarized as [14,15]:
whereA, A’,B, and B’ are fuzzy sets of any kind. This relationship is
expressed as B0 ¼ A0�ðA ! BÞ. Figure 3 shows an example of
Interval Type-2 direct reasoning with Interval Type-2 Fuzzy Inputs.
A fuzzy inference system is a rule base system that uses fuzzy
logic, instead of Boolean logic utilized in data analysis [16,17]. Its
basic structure includes four components (Fig. 4):
� Fuzzifier: Translates inputs (real values) to fuzzy values.� Inference System: Applies a fuzzy reasoning mechanism to
obtain a fuzzy output.� Type Defuzzifer/Reducer: The defuzzifier translates one
output to precise values; the type reducer transforms a
Type-2 Fuzzy Set into a Type-1 Fuzzy Set.
Figure 3 Interval type-2 fuzzy reasoning. [Color figure can be viewed in the online issue, which is available at
wileyonlinelibrary.com.]
4 CASTILLO, MELIN, AND CASTRO
� Knowledge Base: Contains a set of fuzzy rules, and a mem-
bership functions set known as the database.
The decision process is a task that identifies parameters by the
inference system using the rules of the fuzzy rule base [18–20].
These fuzzy rules define the connection between the input
and output fuzzy variables. A fuzzy rule has the form: IF
<Antecedent> THEN <Consequent>, where antecedent is a
composite fuzzy logic expression of one or more simple fuzzy
expressions connected with fuzzy operators; and the consequent
is an expression that assigns fuzzy values to output variables
[21,22].
INTERVAL TYPE-2 FUZZY SYSTEM DESIGN
The Mamdani and Takagi-Sugeno-Kang (TSK) Interval Type-2
Fuzzy Inference Models [23,24] and the design of Interval Type-2
membership functions and operators are implemented in the
IT2FLS Toolbox (Interval Type-2 Fuzzy Logic Systems) reused
from the Matlab1 commercial Fuzzy Logic Toolbox.
The IT2FLS Toolbox includes a series of folders called
dit2mf, it2fis, it2mf, and it2op (Fig. 5). These folders contain
the functions to create Mamdani and TSK Interval Type-2 Fuzzy
Inference Systems (newfistype2.m), functions to add input–output
variables and their ranges (addvartype2.m), it has functions to add
22 types of Interval Type-2 Membership functions for input–
outputs (addmftype2.m), functions to add the rule matrix (addru-
letype2.m), it can evaluate the Interval Type-2 Fuzzy Inference
Systems (evalifistype2.m), evaluate Interval Type-2 Membership
functions (evalimftype2.m), it can generate the initial parameters
of the Interval Type-2Membership functions (igenparamtype2.m),
it can plot the Interval Type-2 Membership functions with the
input–output variables (plotimftype2.m), it can generate the
solution surface of the Fuzzy Inference System (gensurftype2.m),
it plots the Interval Type-2 membership functions (plot2dtype2.m,
plot2dctype2.m), a folder to evaluate the derivatives of the Interval
Type-2Membership Functions (dit2mf) and a folder with different
and generalized Type-2 Fuzzy operators (it2op, t2op).
The Interval Type-2 Fuzzy Inference Systems (IT2FIS) struc-
ture is the MATLAB object that contains all the interval type-2
fuzzy inference system information. This structure is stored inside
each graphics user interface (GUI) tool. Access functions such as
getifistype2 and setifistype2make it easy to examine this structure.
All the information for a given fuzzy inference system is contained
in the IT2FIS structure, including variable names, membership
function definitions, and so on. This structure can itself be thought
of as a hierarchy of structures, as shown in the following diagram
(Fig. 6).
The implementation of the IT2FLS GUI is analogous to the
GUI used for Type-1 FLS in the Matlab1 Fuzzy Logic Toolbox,
thus permitting the experienced user to adapt easily to the use of
IT2FLS GUI. Figures 7 and 8 show the main view of the Interval
Type-2 Fuzzy Systems Structure Editor called IT2FIS (Interval
Type-2 Fuzzy Inference Systems).
SIMULATION RESULTS WITH THE INTERVAL TYPE-2TOOLBOX
We present simulation results for the shower and truck backer–
upper problems with interval type-2 fuzzy logic systems using the
IT2FLS Toolbox. These benchmark problems are available in the
well-known Fuzzy Logic Toolbox ofMatlab (this is a type-1 fuzzy
logic toolbox) and for this reason are a good choice to illustrate the
Figure 4 Type-2 inference fuzzy system structure.
Figure 5 Toolbox folders.
COMPUTATIONAL INTELLIGENCE SOFTWARE 5
results of type-2 fuzzy logic. We also present results with type-2
fuzzy logic for theMackey–Glass time series forecasting, which is
another benchmark problem.
Shower Control Simulation
The application of the interval type-2 fuzzy control scheme to the
shower gives good control results, which can be noticed in
Figure 9.
Truck Backer–Upper Control Simulation
The application of the interval type-2 fuzzy control scheme to the
truck backer–upper problem gives good control results and in
Figures 10 and 11we show the control results of the car trajectories
for different initial conditions.
The results in Table 2 show a comparative analysis for the
Mackey–Glass chaotic time series. The study considers the use of
intelligent hybrid methods, with neural networks (Mamdani,
TSK), type-1 fuzzy inference systems and genetic algorithms
(neuro-genetic, fuzzy-genetic, and neuro-fuzzy) and an interval
type-2 fuzzy logicmodel, for the implicit knowledge acquisition in
a time series behavioral historical data. To identify the model we
use 5 delays, L5x(t), 6 periods, and 500 training data values to
forecast 500 output values. The IT2FLS system works with 4
inputs, 2 interval type-2 Gaussian membership functions for each
input, 16 rules and 1 output with 16 interval linear functions. The
forecasted root mean square error (RMSE) is 0.0335 (Table 2).
In Table 2, NNFF stands for a neuro-fuzzymodel, CANFIS is
a modified adaptive neuro-fuzzy inference system, NNFF-GA
stands for genetically optimized neuro-fuzzy system,
FLS(TSK)-GA means a Sugeno fuzzy model optimized with a
genetic algorithm, FLS(MAM)-GA means a Mamdani fuzzy
model optimized with a genetic algorithm, and IT2FLS an interval
type-2 fuzzy model developed with the type-2 fuzzy logic toolbox
presented in this article.
Figure 6 Hierarchy of IT2FIS structures diagram.
6 CASTILLO, MELIN, AND CASTRO
RELATED WORK
In comparing the developed toolbox (shown as IT2tool in Table 3)
with available software for type-2 and type-1 fuzzy systems in the
literature [25,26],we can find that no other tool exists for designing
and implementing hybrid type-2 fuzzy systems (like neuro-type-2
fuzzy systems). Other tools could be used for the same task, but a
lot of programming would be needed by the user. In Table 3, a
summary of a comparison of the IT2tool, shown in the last column,
with respect to other tools is presented. The comparison is made
based on the main characteristics needed of a good software tool
for designing and implementing type-2 fuzzy systems, this is
indicated in the first column of Table 3. It is easy to notice that
only the proposed IT2tool fulfills all the requirements of an ideal
tool for this purpose. Other software tools that only have some of
the ideal characteristics are not flexible enough for complex and
high-dimensional problems.
There are other educational software tools available for fuzzy
logic [30], but in most of the cases the software is only for type-1
fuzzy logic or is focused on a particular area of application. The
proposed type-2 fuzzy logic toolbox is for any type of application
and can be used for type-1 or type-2 fuzzy logic.
VALIDATION OF THE SOFTWARE
The validation of the developed software has been done by com-
paring the results of the toolbox against the results of other authors
with the same benchmark problems [31–33]. The type-2 fuzzy
logic results have been compared in several papers by the authors
Figure 7 IT2FIS Editor. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Figure 8 Interval Type-2 MF’s Editor. [Color figure can be viewed in the online issue, which is available at
wileyonlinelibrary.com.]
COMPUTATIONAL INTELLIGENCE SOFTWARE 7
[34–38]. In fact, for this reason the type-2 fuzzy logic toolbox has
been requested by many researchers and professors of different
countries, interested in teaching or doing research with the toolbox.
On other hand, with respect to the education process, we can
say that the toolbox has been in use for the last four semesters (of
2009 and 2010) with students of a graduate course at the Masters
Degree in Computer Science with great success. The graduate
course is called ‘‘Fuzzy Systems’’ and a total 100 students have
used the toolbox with great success. This course is a basic part of
the Master degree in Computer Science at Tijuana Institute of
Technology because the area of research of the graduate program is
on Computational Intelligence.
With the goal of measuring the performance of the proposed
software for type-2 fuzzy logic (IT2tool) a questionnaire was
Figure 9 Temperature interval type-2 fuzzy control. [Color figure can be viewed in the online issue, which is available at
wileyonlinelibrary.com.]
Figure 10 Interval type-2 fuzzy control for trajectory 1. [Color figure can be viewed in the online issue, which is
available at wileyonlinelibrary.com.]
8 CASTILLO, MELIN, AND CASTRO
applied to a sample of n ¼ 30 graduate students that have used the
tool for a semester in a course of fuzzy systems. The students were
also providedwith the Fuzzy Logic toolbox ofMatlab, so that they
could be able to compare the characteristics and performance of
both tools in solving problems using fuzzy logic. The results of the
questionnaire are summarized in Table 4, where the nine questions
are indicated in the first column of this table. The students were
asked to provide a ranking from 1 to 10 to the response to each of the
questions. The averages of questions’ responses of the 30 students
for each of the tools are indicated in the second and third columns
of Table 4, respectively. The summary of the statistics (average and
standard deviations) is indicated at the bottom of Table 4.
The t-statistic value of 4.75 for the difference ofmeans is also
shown at the bottom of Table 4, which indicates that there is
significant difference between the results of our software tool
(IT2tool) those ones provided by Matlab with at least 99% con-
fidence. In other words, the user’s satisfaction with the developed
tool (IT2tool) is significantly better than with the Fuzzy
Logic toolbox of Matlab. Finally, from Table 4 we can notice that
in all the questions the response was in favor (on average) of the
developed tool (IT2tool).
Finally, for measuring the knowledge improvement of the
students when using the software tool, we use a metric based on
two factors: accuracy of the student’s achieved solutions and time
required by the student to develop the solutions to the problem. The
Figure 11 Interval type-2 fuzzy control for trajectory 2. [Color figure can be viewed in the online issue, which is
available at wileyonlinelibrary.com.]
Table 2 Forecasting of Time Series
Methods
Mackey–Glass
RMSE trn/chk epochs cpu(s)
NNFF 0.0595 500/500 200 13.36
CANFIS 0.0016 500/500 50 7.34
NNFF-GA 0.0236 500/500 150 98.23
FLS(TKS)-GA 0.0647 500/500 200 112.01
FLS(MAM)-GA 0.0693 500/500 200 123.21
IT2FLS 0.0335 500/500 6 20.23
Table 3 Comparison With Different Fuzzy Logic Software
Characteristics
Mendel
[27]
Ozek
Toolbox [28]
Matlab
[29] IT2tool
Graphic interface No Yes Yes Yes
Hybrid algorithms No No Yes Yes
Type-2 Fuzzy Logic Yes Yes No Yes
Neuro-Fuzzy Type-2 No No No Yes
Easy to use No Yes Yes Yes
Efficiency (in time) No Yes Yes Yes
Easy to add new methods Yes Yes No Yes
Quick to shows results No No No Yes
Table 4 Summary of the Questionnaire Applied to Students
Question IT2tool Matlab Fuzzy
User interface 9.2560 8.8221
Method for showing results 9.3333 8.7350
Method to input data 9.1251 8.6725
Efficiency (in time) 9.0550 8.4322
Accuracy 9.5275 9.0531
Comparison between both tools 9.4188 8.6337
Comparison with other software 9.6713 8.9572
Global evaluation 9.9071 9.2879
Recommending the tool to others 9.2781 8.6841
Average 9.3970 8.8090
Standard deviation 0.2700 0.2560
t Statistic 4.75
COMPUTATIONAL INTELLIGENCE SOFTWARE 9
metric can be expressed as given by the following equation:
K ¼ aAþ bT (7)
where K is the knowledge of the student, A is the accuracy of the
achieved solution by the student t, T is the time required by the
student to develop the solution to the problem. In this equation, all
variables are considered to be normalized in the range from0 to 10,
to represent an expert evaluation of the achieved result by the
student. The coefficients a and b represent the weights assigned to
each of the factors involved in evaluating the knowledge acquired
by the student. In Table 5 the results achieved by students before
and after learning to use the software tool are presented. These
results were obtained by giving 30 students the same five bench-
mark problems to solve. Two of the problems are of time series
prediction (Mackey–Glass and Dow Jones time series). The other
three problems are of intelligent control (Shower Control, Truck
Backer–upper control, and Ball and Beam control). From the
statistics of Table 5, in particular the t value, we can conclude
that there exists statistical evidence of significant difference bet-
ween the average knowledge before and after learning to use the
software tool proposed in this paper. In other words, the software
tool provides significant additional knowledge to the students
based on this statistical evidence. This conclusion is achieved
with more than 98% of confidence because the t value if of 3.11.
CONCLUSIONS
The results in the interval type-2 fuzzy control case studies of the
shower and the truck backer–upper have similar results to the type-
1 fuzzy control with moderate uncertain footprints. To better
characterize the interval type-2 fuzzy models we need to generate
more case studies with better knowledge bases for the proposed
problems, therefore classifying the interval type-2 fuzzy model
application strengths.
The design and implementation done in the IT2FLS Toolbox
are potentially important for research in the interval type-2 fuzzy
logic area, thus solving complex problems on the different applied
areas.
The main contribution of the article has been to facilitate the
use of the theory and concepts of type-2 fuzzy logic for both
researchers and students. On the educational side there has been a
significant improvement in the level of knowledge of the students
that have used the toolbox, which is appreciated with a performed
evaluation to a sample of students.
Our future work is to improve the IT2FLS Toolbox with a
better GUI, to have compiled code, and integrate a learning
technique Toolbox to optimize the knowledge base parameters
of the interval type-2 fuzzy inference system, and design interval
type-2 fuzzy neural network hybrid models.
REFERENCES
[1] L. A. Zadeh, Fuzzy sets, Inf Control 8 (1965), 338–353.
[2] L. A. Zadeh, Outline of a new approach to the analysis of complex
systems and decision processes, IEEE Trans Syst Man Cybern 3
(1973), 28–44.
[3] L. A. Zadeh, The concept of a linguistic variable and its applica-
tion to approximate reasoning, parts 1, Inf Sci 8 (1975), 199–249.
[4] L. A. Zadeh, Fuzzy logic, Computer 1 (1988), 83–93.
[5] L. A. Zadeh, Knowledge representation in fuzzy logic, IEEE Trans
Knowl Data Eng 1 (1989), 89–100.
[6] L. A. Zadeh, Fuzzy logic ¼ computing with words, IEEE Trans
Fuzzy Syst 2 (1996), 103–111.
[7] N. N. Karnik and J. M. Mendel, An introduction to type-2 fuzzy
logic systems, University of Southern California, Los Angeles, CA,
1998.
[8] Q. Liang and J. Mendel, Interval type-2 fuzzy logic systems: Theory
and design, IEEE Trans Fuzzy Syst 8 (2000), 535–550.
[9] O.Castillo andP.Melin, Type-2 fuzzy logic: Theory and applications,
Springer-Verlag, Heidelberg, Germany, 2008.
[10] J.M.Mendel, Uncertain rule-based fuzzy logic systems: Introduction
and new directions, Prentice-Hall, NJ, 2001.
[11] M.Mizumoto andK. Tanaka, Some properties of fuzzy sets of type-2,
Inf Control 31 (1976), 312–340.
[12] L.-X. Wang, Adaptive fuzzy systems and control: Design and
stability analysis, Prentice Hall, Englewood Cliffs, NJ, 1994.
[13] R. R. Yager, On the implication operator in fuzzy logic, Inf Sci 31
(1983), 141–164.
[14] R. R. Yager, A characterization of the fuzzy extension principle,
Fuzzy Sets Syst 18 (1986), 205–217.
[15] E. H.Mamdani and S. Assilian, An experiment in linguistic synthesis
with a fuzzy logic controller, Int J Man Mach Stud 7 (1975), 1–13.
[16] E. H. Mamdani, Applications of fuzzy logic to approximate reasoning
using linguistic synthesis, IEEE Trans Comput 26 (1977), 1182–1191.
[17] M. Sugeno and G. T. Kang, Structure identification of fuzzy model,
Fuzzy Sets Syst 28 (1988), 15–33.
[18] M. Sugeno, Industrial applications of fuzzy control, Elsevier Science,
Amsterdam, Holland, 1985.
[19] Y. Tsukamoto, An approach to fuzzy reasoning method. In: M. M.
Gupta, R. K. Ragade, R. R. Yager (Eds.), Advances in fuzzy set
theory and applications, Elsevier, North-Holland, Amsterdam, 1979,
pp 137–149.
[20] R. R. Yager, On a general class of fuzzy connectives, Fuzzy Sets Syst
4 (1980), 235–242.
[21] E. H. Mamdani, Advances in the linguistic synthesis of fuzzy con-
trollers, Int J Man Mach Stud 8 (1976), 669–678.
Table 5 Experimental Results of Evaluating Knowledge Improvement With the Software Tool
Benchmark problem
Knowledge improvement
Average accuracy
(before)
Average time
(before)
K
(before)
Average accuracy
(after)
Average time
(after)
K
(after)
Mackey–Glass time series 8.15 8.33 8.24 9.55 9.60 9.575
Dow Jones time series 7.20 7.84 7.52 8.72 8.96 8.84
Shower control 7.45 8.25 7.85 9.15 9.35 9.25
Truck Backer–upper control 9.25 9.15 9.20 9.72 9.84 9.78
Ball and Beam control 6.95 7.47 7.21 8.24 9.28 8.76
Average 8.004 Average 9.241
Standard deviation 0.770 Standard deviation 0.446
Statistic t ¼ 3.11
10 CASTILLO, MELIN, AND CASTRO
[22] T. Takagi and M. Sugeno, Fuzzy identification of systems and its
applications to modeling and control, IEEE Trans Syst Man Cybern
15 (1985), 116–132.
[23] J. R. Castro, O. Castillo, P. Melin, and A. Rodriguez-Diaz, Building
fuzzy inference systems with a new interval type-2 fuzzy logic
toolbox, Trans Comput Sci 1 (2008), 104–114.
[24] R. Sepulveda, O. Castillo, P. Melin, A. Rodriguez-Diaz, and O.
Montiel, Experimental study of intelligent controllers under uncer-
tainty using type-1 and type-2 fuzzy logic, Inf Sci 177 (2007), 2023–
2048.
[25] J. R. Castro, O. Castillo, and P. Melin, An interval type-2 fuzzy logic
toolbox for control applications, Proceedings of FUZZ-IEEE,
London, 2007, pp 1–6.
[26] P. Melin and O. Castillo, Hybrid intelligent systems for pattern
recognition, Springer-Verlag, Heidelberg, Germany, 2005.
[27] J. M. Mendel, Type-2 fuzzy logic software, available at: http://sipi2-
b.usc.edu/�mendel/software/.
[28] M. B. Ozek and Z. H. Akpolat, A software tool: Type-2 fuzzy logic
toolbox, Comput Appl Eng Educ 16 (2008), 137–146.
[29] Logic. Fuzzy, Toolbox of Matlab, available at: http://www.math-
works.com/products/fuzzylogic/.
[30] M. A. Akcayol, C. Elmas, O. A. Erdem, and M. Kurt, An educational
tool for fuzzy logic controller and classical controllers, Comput Appl
Eng Educ 12 (2004), 126–135.
[31] O. Mendoza, P. Melin, and G. Licea, Fuzzy inference systems type-1
and type-2 for digital images edges detection, EngLett 15 (2007), 45–
52.
[32] O. Mendoza, P. Melin, and G. Licea, Interval type-2 fuzzy logic for
edges detection in digital images, Int J Intell Syst 24 (2009), 1115–
1133.
[33] C. Leal-Ramirez, O. Castillo, P. Melin, and A. Rodriguez-Diaz,
Simulation of the bird age-structured population growth based on
an interval type-2 fuzzy cellular structure, Inf Sci 181 (2011), 519–
535.
[34] N. R. Cazarez-Castro, L. T. Aguilar, and O. Castillo, Fuzzy logic
control with genetic membership function parameters optimization
for the output regulation of a servomechanism with nonlinear back-
lash, Expert Syst Appl 37 (2010), 4368–4378.
[35] J. R. Castro, O. Castillo, P. Melin, and A. Rodriguez-Diaz, A hybrid
learning algorithm for a class of interval type-2 fuzzy neural net-
works, Inf Sci 179 (2009), 2175–2193.
[36] R. Martinez-Soto, O. Castillo, and L. T. Aguilar, Optimization of
interval type-2 fuzzy logic controllers for a perturbed autonomous
wheeled mobile robot using genetic algorithms, Inf Sci 179 (2009),
2158–2174.
[37] O. Castillo, L. T. Aguilar, N. R. Cazarez, and S. Cardenas, Systematic
design of a stable type-2 fuzzy logic controller, Appl Soft Comput 8
(2008), 1274–1279.
[38] R. Sepulveda, O. Castillo, P. Melin, and O. Montiel, An
efficient computational method to implement type-2 fuzzy logic in
control applications, analysis and design of intelligent systems using
soft computing techniques 2007, Springer-Verlag, Heidelberg,
Germany, pp 45–52.
BIOGRAPHIES
Oscar Castillo holds the Doctor in Science
degree (Doctor Habilitatus) in Computer
Science from the Polish Academy of Sciences
(with the Dissertation ‘‘Soft Computing and
Fractal Theory for Intelligent Manufacturing’’).
He is a Professor of Computer Science in
the Graduate Division, Tijuana Institute of
Technology, Tijuana, Mexico. In addition, he
is serving as Research Director of Computer
Science and head of the research group on
fuzzy logic and genetic algorithms. Currently, he is Vice-President of
HAFSA (Hispanic American Fuzzy Systems Association) and President
Elect of IFSA (International Fuzzy Systems Association). Prof. Castillo is
also Chair of the Mexican Chapter of the Computational Intelligence
Society (IEEE). He belongs to the Mexican Research System with level
II. He belongs to the Technical Committee on Fuzzy Systems of IEEE,
where he is the Chair of the Task Force on FuzzyHardware and also belongs
to the Task Force on ‘‘Extensions to Type-1 Fuzzy Systems’’. He is also a
member of NAFIPS, IFSA and IEEE. His research interests are in Type-2
Fuzzy Logic, Fuzzy Control, Neuro-Fuzzy and Genetic-Fuzzy hybrid
approaches. He has published over 90 journal papers, 5 authored books,
20 edited books, and 200 papers in conference proceedings.
Patricia Melin holds the Doctor in Science
degree (Doctor Habilitatus D.Sc.) in Computer
Science from the Polish Academy of Sciences
(with the Dissertation ‘‘Hybrid Intelligent
Systems for Pattern Recognition using Soft
Computing’’). She is a Professor of Computer
Science in the Graduate Division, Tijuana Insti-
tute of Technology, Tijuana, Mexico, since
1998. In addition, she is serving as Director of
Graduate Studies in Computer Science and head
of the research group on Computational Intelligence (2000-present).
Currently, she is President of HAFSA (Hispanic American Fuzzy Systems
Association and is the founding Chair of the Mexican Chapter of the IEEE
Computational Intelligence Society. She is Chair of the Task Force on
Hybrid Intelligent Systems ofNeuralNetworks Technical Committee of the
IEEE Computational Intelligence Society. She is member of NAFIPS,
IFSA, and IEEE. She belongs to the Mexican Research System with level
II. Her research interests are in Type-2 Fuzzy Logic, Modular Neural
Networks, Pattern Recognition, Neuro-Fuzzy and Genetic-Fuzzy hybrid
approaches. She has published over 90 journal papers, 5 authored books, 15
edited books, and 200 papers in conference proceedings.
Juan R. Castro holds the Ph.D. in
Computer Science from UABC University in
Tijuana, Mexico (2009) and the Masters
Degree in Computer Science from Tijuana
Institute of Technology (2005). Currently he
is Professor of Computer Science in the School
of Engineering of UABC University, where he
lectures at the undergraduate and graduate
levels. His research areas interests are in
type-2 fuzzy logic, genetic fuzzy systems,
type-2 neuro-fuzzy neural networks, and new techniques for time series
prediction and intelligent control.
COMPUTATIONAL INTELLIGENCE SOFTWARE 11