!

Err TRANSACTIONS O N

SYSTEMS, MAN. ANI) CYHT.RNT.TI(S, VOI..

20, NO. 2.

MAR(.II/APRII.

1000

419

Fuzzy Logic in Control Systems: Fuzzy Logic Controller, Part I1

Abstract -During the past several years, fuzzy control has emerged as one of the most active and fruitful areas for research in the applications of fuzzy set theory, especially in the realm of industrial processes, which do not lend themselves to control by conventional methods because of a lack of quantitative data regarding the input-output relations. Fuzzy control is based on fuzzy logic-a logical system that is much closer in spirit to human thinking and natural language than traditional logical systems. The fuzzy logic controller (FLC) based on fuzzy logic provides a means of converting a linguistic control strategy based on expert knowledge into an automatic control strategy. A survey of the FLC is presented; a general methodology for constructing an FLC and assessing its performance is described; and problems that need further research are pointed out. In particular, the exposition includes a discussion of fuzzification and defuzzification strategies, the derivation of the database and fuzzy control rules, the definition of fuzzy implication, and an analysis of fuzzy reasoning mechanisms.

I. DECISIONMAKING LOGIC

S WAS noted in Part I of this paper [150], an FLC may be regarded as a means of emulating a skilled human operator. More generally, the use of an FLC may be viewed as still another step in the direction of modeling human decisionmaking within the conceptual framework of fuzzy logic and approximate reasoning. In this context, the forward data-driven inference (generalized modus ponens) plays an especially important role. In what follows, we shall investigate fuzzy implication functions, the sentence connectives and and also, compositional operators, inference mechanisms, and other concepts that are closely related to the decisionmaking logic of an FLC.

I ) Basic Properties of a Fuuy Implication Function: The choice of a fuzzy implication function involves a number of criteria, which are discussed in [31, [241, [2l, [711, [181, [521, [191, [116], [ S I , [72], and [96]. In particular, Baldwin and Pilsworth [3] considered the following basic characteristics of a fuzzy implication function: fundamental property, smoothness property, unrestricted inference, symmetry of generalized modus ponens and generalized modus tollens, and a measure of propagation of fuzziness. All of these properties are justified on purely intuitive grounds. We prefer to say that the inference (consequence) should be as close as possible to the input truth function value, rather than be equal to it. This gives us a more flexible criterion for choosing a fuzzy implication function. Furthermore, in a chain of implications, it is necessary to consider the fuzzy syllogism [ 1471 associated with each fuzzy implication function .before we can talk about the propagation of fuzziness. Fukami, Mizumoto, and Tanaka [24] have proposed a set of intuitive criteria for choosing a fuzzy implication function that constrains the relations between the antecedents and consequents of a conditional proposition, with the latter playing the role of a premise in approximate reasoning. As is well known, there are two important fuzzy implication inference rules in approximate reasoning. They are the generalized modus ponens (GMP) and the generalized modus tollens (GMT). Specifically,
premise 1: x is A

A. Fuzzy Implication Functions

In general, a fuzzy control rule is a fuzzy relation which is expressed as a fuzzy implication. In fuzzy logic, there are many ways in which a fuzzy implication may be defined. The definition of a fuzzy implication may be expressed as a fuzzy implication function. The choice of a fuzzy implication function reflects not only the intuitive criteria for implication but also the effect of connective also.
Manuscript received May 27, 198X; revised July 1, 19x9. This work wpported in part by NASA under grant NCC-2-275 and in part by AFOSR under grant 89-0084. The author is with the Electronics Research Laboratory, Department of Electrical Engineering and Computer Sciences, University of California, Berkeley, CA 94720. IEEE Log Number X032014.

Dremise 2: if x is A then y is B consequence: y is B premise 1: y is B oremise 2: if x is A then v is B consequence: x is A

(GMP)

(GMT)

in which A , A, B , and B are fuzzy predicates. The propositions above the line are the premises; and the proposition below the line is the consequence. The proposed criteria are summarized in Tables I and 11. We note that if a causal relation between x is A and y is B is not strong in a fuzzy implication, the satisfaction of criterion 2-2 and criterion 3-2 is allowed. Criterion 4-2 is interpreted as: if x is A then y is B , else y is not B.

0018-9472/90/0300-0419$01.00 01990 IEEE

420

IEEE T R A N S A C T I O N S O N

SYSTTMS. M A N ,

ANI) ('YBERNETI('S.

Vol..

20.

NO.

2,

MAlZ('II/Al'KII.

1990

TABLE I INTUITIVECRITEKIA RTLATING PREI AND CONS FOR GIVEN PRE2 IN GMP
x is A'(Pre1)

I N T U I T I V E CRITERIA

FOR

TABLE 11 REIATING PRF.1 ANI) CONS GIVEN P R EI~ N GMT

x is AYCons)

y is BYCons) y is B y is very B y is B y is more or less B y is B y is unknown y is not B

Criterion 5 Criterion 6 Criterion 7 Criterion 8-1 Criterion 8-2

v is B'(Pre1) y is not B y is not very B y is not more or less B y is B y is B

x is

Criterion 1 Criterion 2-1 Criterion 2-2 Criterion 3-1 Criterion 3-2 Criterion 4-1 Criterion 4-2

x is A
x is very A

x is very A
x is more or less A

not A x is not very A x is not more or less A

x is unknown x is A

x is more or less A x is not A x is not A

defined for all x , y union

E [O,

11:
x

vy

= max{x, y )

Although this relation is not valid in formal logic, we often make such an interpretation in everyday reasoning. The same applies to criterion 8. 2) Families of Fuzzy Implication Functions: Following Zadeh's [146] introduction of the compositional rule of inference in approximate reasoning, a number of researchers have proposed various implication functions in which the antecedents and consequents contain fuzzy variables. Indeed, nearly 40 distinct fuzzy implication functions have been described in the literature. In general, they can be classified into three main categories: the fuzzy conjunction, the fuzzy disjunction, and the fuzzy implication. The former two bear a close relation to a fuzzy Cartesian product. The latter is a generalization of implication in multiple-valued logic and relates to the extension of material implication, implication in propositional calculus, modus ponens, and modus tollens [18]. In what follows, after a short review of triangular norms and triangular co-norms, we shall give the definitions of fuzzy conjunction, fuzzy disjunction, and fuzzy implication. Some fuzzy implication functions, which are often employed in an FLC and are commonly found in the literature, will be derived. Definition 1: Triangular Noms: The triangular norm * is a two-place function from [O, 11 x [O, 11 to [O, 11, i.e., * : [O, 11 X [0, 13 + [O, 11, which includes intersection, algebraic product, bounded product, and drastic product. The greatest triangular norm is the intersection and the least one is the drastic product. The operations associated with triangular norms are defined for all x , y E [0,1]: intersection algebraic product bounded product drastic product
x Ay
x.y
=

algebraic sum bounded sum drastic sum disjoint sum

x*y=x+y-xy
xey
= min{l,x

+y)

xuy= y 1

i
x

y=o x=o x,y>o

x A y = max(min(x, 1 - y ) , min(1- x , y ) } .

The triangular norms are employed for defining conjunctions in approximate reasoning, while the triangular conorms serve the same role for disjunctions. A fuzzy control rule, "if x is A then y is B," is represented by a fuzzy implication function and is denoted by A + B, where A and B are fuzzy sets in universes U and V with membership functions p A and p B , respectively. Definition 3: Fuuy Conjunction: The fuzzy conjunction is defined for all U E U and v E V by A+B=AXB

=/I,x~pa(u)*pB(v)/(u,v)
where * is an operator representing a triangular norm. Definition 4: Fuuy Disjunction: The fuzzy disjunction is defined Cor all U E U and v E V by A+B=AXB

min ( x , y )

xy
max(0,x

xOy

+y

1)

where + is an operator representing a triangular co-norm. Definition 5: Fuzzy Implication: The fuzzy implication is associated with five families of fuzzy implication functions is a in use. As before, * denotes a triangular norm and triangular co-norm.

xny=

i
x y

y=l x=l x,y<l.

4.1) Material implication: A + B =(not A ) + B 4.2) Propositional calculus:

A --+ B = (not A ) + ( A * B ) 4.3) Extended propositional calculus:
A -+ B = (not A x not B ) + B 4.4) Generalization of modus ponens: A 4 B = s u p { c E [OJ], A * c < B }

Definition 2: Triangular CO-Noms: The triangular conorms + is a two-place function from [O, l ] x [O, l ] to [O, l], i.e. + : [O, 11 X [O, 11 to [O, 11, which includes union, algebraic sum, bounded sum, drastic sum, and disjoint sum. The operations associated with triangular co-norms are

LEE: FUZZY LOGIC' IN ('ONTROI. SYSTEMS: FUZZY I . O C i l ( ' ('0NTKOI.I.TK-I'ART

II

42 1

from Definition 5.1 by using the union; and Goguen's fuzzy implication follows from Definition 5.4 by using the A+B=inf{t~[O,l],B + t < A } algebraic product. 3) Choice of a Fuzzy Implication Function: First, we Based on these definitions, many fuzzy implication funcinvestigate the consequences resulting from applying the tions may be generated by employing the triangular norms preceding forms of fuzzy implication in fuzzy inference and co-norms. For example, by using the definition of the and, in particular, the GMP and GMT. The inference is fuzzy conjunction, Mamdani's mini-fuzzy implication, R c , based on the sup-min compositional rule of inference. In is obtained if the intersection operator is used. Larsen's the GMP, we examine the consequence of the following product fuzzy implication, R,, is obtained if the algebraic compositional equation: product is used. Furthermore, R,, and R,, are obtained B ' = A'o R if the bounded product and the drastic product are used, respectively. The following fuzzy implications, which are where often adopted in an FLC, will be discussed in more detail R fuzzy implication (relation), at a later point. sup-min compositional operator, A' a fuzzy set which has the form: Mini-operation rule of fuzzy implication [Mamdani]: A = /upA(u)/u R,=AXB very A = A 2 = j U p ; ( u ) / u more or less A = A".' = /up'j45(u)/u not A = jU1 -pA(u)/u. Product operation rule of fuzzy implication [Larsen]:

4.5) Generalization of modus tollens:

R,=AxB
P A ( U ) P . B ( V ) / ( U , v). /"xv Arithmetic rule of fuzzy implication [Zadeh]: R , =(not A x V ) @ ( UX B )

Similarly, in the GMT, we examine the consequence of the fcllowing equation: A'=RoB' where

Maxmin rule of fuzzy implication [Zadeh]: R , = ( A x B)u(not A x V )

pJv))'J(1~A(u))/(u,v).

R fuzzy implication (relation) B' a fuzzy set that has the form: not B = /,,1- p B ( u ) / u not very B = jvl- p i ( v ) / v not more or less B = /,,1- p 3 v ) / v B = j v p B ( v ) / v.
The Case of R,: Larsen's Product Rule: A method for computing the generalized modus ponens and the generalized modus tollens laws of inference is described in [3]. The graphs corresponding to Larsen's fuzzy implication R, are given in Fig. 1. The graph with parameter pA is used for the GMP, and the graph with p B is used for the GMT. Larsen's Product Rule in GMP: Suppose that A' = A* (a> 0); then the consequence B; is inferred as follows:

Standard sequence fuzzy implication: R,= A x V + U X B

> pJv))/(u, v)
where

Boolean fuzzy implication: R , =(not A x V ) U ( U x B )

p.A(u))v(pU,(v))/(u,v).

B;

= A"

R,,

Goguen's fuzzy implication: RA=A x V + U X B

> > p.B(v>)/(u, v)

where
PA(U)>>

The membership function pBL of the fuzzy set B; is pointwise defined for all v E V by
p B; (

pJv)=

We note that Zadeh's arithmetic rule follows from Definition 5.1 by using the bounded sum operator; Zadeh's maxmin rule follows from Definition 5.2 by using the intersection and union operators; the standard sequence implication follows from Definition 5.4 by using the bounded product; Boolean fuzzy implication follows

i '
CLB(U)

CLA(U) < r U B ( V )

) = sup min ( ll"A ( ) PA( ) pf3(

9

PAU)

>P B ( 4 .
=

11 E

U U

SUP q 1 - P > < 4 >

I1 E

PA(V)

where
qIL%))

A min I

P X 4 PA(4ILB(4).
9

{A' = A): The values of Sp(pA(u))with a parameter pB(v),say p B ( v )= 0.3 and 0.8, are indicated in Fig. 2 by a
broken line and dotted line, respectively. The member-

422

S
1.o 1 .o

0.5

0.5

0 0
Fig. 3. Approximate reasoning: generalized modus ponens with Larsen's product operation rule.

(b) Fig. 1. Diagrams for calculation of membership functions. (a) flR ~ pR,, . versus pB with parametef versus pAwith the parameter j ~ (b) PA.

0
Fig. 4. Approximate reasoning: generalized modus ponens with Larsen's product operation rule.

sp(lgA)

1-BA(()
0.5

.
AA(()

.?'.

-------.'\ 0.3''\.\.

Fig. 2.

Approximate reasoning: generalized modus ponens with Larsen's product operation rule.

ship function pBLis obtained by CLB$V)

= =

Fig. 5. Approximate reasoning: generalized modus ponens with Larsen's product operation rule.
9

SUP min (PA( U )

UU UU

PA( U )P B (

V) I

SUP PA(U)CLB(V)

by a broken line and dotted line, respectively. The membership function pnL is given by PB;,(V) = SUP min{l.~"(.),pA(u),H(V~l
U E

CLA(U)= I =pn(V), {AA'= A2): The values of S,,(p:(u)) with a parameter pB(v), say pB(v) = 0.3 and 0.8, are indicated in Fig. 3 by a broken line and dotted line, respectively. The membership function psi may be expressed as

ti

=pn(

v).

{A' = not A): The values of S,( - p A ( u ) ) with a pa-

PB$V)

SUP min(Cl~(u),CLA(u)pB(v)J
UU

rameter p s ( ~ ) ,say p s ( ~ ) = 0.3 and 0.8, are indicated in Fig. 5 by a broken line and dotted line, respectively. The membership function psh is given by
P BP, ( v ) = SUP m i n I l - p A ( U ) , p A ( U ) / * H ( v ) l
U E

=PB(V). {A' = The values of S,,(py(u)) with a parameter pB(v), say p J v ) = 0.3 and 0.8, are indicated in Fig. 4

ti

P AV I

LEE: FUZZY LOGIC' IN C'ONTKOL SYSTEMS: FUZZY L O G I C ' CONTROI.I.EK-PAKT

II

423

NdV)
Fig. 8. Approximate reasoning: generalized modus tollens withlarsen's product operation rule.

Fig. 6. Approximate reasoning: generalized modus tollens with Larsen's product operation rule.

S
1.o 1.o

0.6

0.5

0
0
1

NdV)

Fig. 7. Approximate reasoning: generalized modus tollens with Larsen's product operation rule.

Fig. 9. Approximate reasoning: generalized modus tollens with Larsen's product operation rule.

Larsen's Product Rule in GMT: Suppose that B'= not B" (a> 0); then the consequence A; is inferred as follows:
A; = R, o(not B")

TABLE 111

SUMMARY OF INFERENCE RESULTS FOR GENERALIZED MODUSPONENS

A

Very A

More or Less A
PB
PB

Not A 0.5 A
PB l+PB

4 s - 1
, l L

1 1 1

The membership function pA; of the fuzzy set A; is pointwise defined for all U E U by
pA;( U ) = sup min { p A (u ) p B (v), 1VV

T 2

6-1
V

B B

fi-1
V P

v)}

SUP
VV

w-

P%V>>

where

s,(1- Prig( V 1)

min I PA( U )PB( V ) > 1- cL% v ) 1 .

{E=not B } : The values of S,(1- p B ( v ) )with a parameter pA(u), say pA(u)= 0.3 and 0.8, are indicated in Fig. 6 by a broken line and dotted line, respectively. The membership function pA; is given by
~,g;(u= ) S~P~~~{CLA(U)I~B(V),~-~B(V)}
VV

Fig. 7 by a broken line and dotted line, respectively. The membership function p A ; is given by

PA( U )
'

{B'= not B2}: The values of S,(1- &(v)) with a parameter pA(u),say p A ( u )= 0.3 and 0.8, are indicated in

The values of S,(1- &j5(v)) with a {B'= not parameter pA(u),say pA(u)= 0.3 and 0.8, are indicated in Fig. 8 by a broken line and dotted line, respectively. The

424

I E E E TRANSA(.TIONS ON SYSTEMS, MAN, A N I ) (YUERNETICS. VOI..

20, NO. 2, M A R ( I I / A P R I L . 1990

TABLE IV SUMMARY OF INFERENCE RESULTS FOR GENERALIZED Moous TOI.I.F.NS

Not B Not Very B Not More or Less B

R,

.FA)

TABLE V SATISFACTION OF VARIOUS Fuzzy IMPLICATION FUNCTIONS UNDER INTUITIVECRITERIA

R,
Criteria 1 Criteria 2-1 Criteria 2-2 Criteria 3-1 Criteria 3-2 Criteria 4-1 Criteria 4-2 Criteria 5 Criteria 6 Criteria 7 Criteria 8-1 Criteria 8-2

R,

R,,
X X X X X

RP>?
X X X X X

R,

Rh
X X X X X

RI,
X X X X X

0
X

0
X

0 0
X

0
X

0
X

0
X

0
X X X X X X

0
X X
X X X X

0
X X X X

0
X
X X X X X

0
X

0
X
X X X

0
X X X X

0
X

0 0 0 0
X

0
X

0
X

membership function p A ; ,is given by

PLA;(U) =

SUP m i n { P A ( U ) P B ( v ) , 1 - P ~ 5 ( V ) }
V V

- 2 P A U ) + 1-

2PA(U)

{B= B): The values of S , ( p B ( v ) )with a parameter

pA(u),say p A ( u )= 0.3 and 0.8, are indicated in Fig. 9 by a

broken line and dotted line, respectively. The membership function pA; is given by
P A ; W = SUP I P A ( U ) P B ( V ) ? P B ( V ) l
V V

=PA(U).

The remaining consequences [24] inferred by R,, R,, R,,R,,R,,R, can be obtained by the same method as just described. The results are summarized in Tables I11 and IV. By employing the intuitive criteria in Tables I and I1 in Tables 111 and IV, we can determine how well a fuzzy implication function satisfies them. This information is summarized in Table V. In FLC applications, a control action is determined by the observed inputs and the control rules, without the

consequent of one rule serving as the antecedent of another. In effect, the FLC functions as a one-level forward data-driven inference (GMP). Thus the backward goal-driven inference (GMT), chaining inference mechanisms (syllogisms), and contraposition do not play a role in the FLC, since there is no need to infer a fuzzy control action through the use of these inference mechanisms. Although R , and R , do not have a well-defined logical structure, the results tabulated in Table V indicate that they are well suited for approximate reasoning, especially for the generalized modus ponens. R , has a logical structure which is similar to R,. R , is based on the implication rule in Lukasiewiczs logic LAleph. However, R , and R , are not well suited for approximate reasoning since the inferred consequences do not always fit our intuition. Furthermore, for multiple-valued logical systems, R , and R A have significant shortcomings. Overall, R, yields reasonable results and thus constitutes an appropriate choice for use in approximate reasoning.

B. Interpretation of Sentence Connectives and, also

In most of the existing FLCs, the sentence connective and is usually implemented as a fuzzy conjunction in a Cartesian product space in which the underlying variables

take values in different universes of discourse. As an illustration, in if ( A and B ) then C, the antecedent is interpreted as a fuzzy set in the product space U X V , with the membership function given by
PA X B ( U ,

SLJlTAH1.F PAIRS OF A

TABLE VI F(JZZY IMP1 I(ATI0N FUNCTION Connective Also

A N D CONNECTIVE LllSO

.>
v)

Implication Rule

= min { P A ( U ) ? P A

4 1
I t depends on the shape of reproduced curve which forms the set of fuzzy control rules.

or
PA x n (
U2

= PA( U

1. P A v 1

where U and V are the universes of discourse associated with A and B , respectively. When a fuzzy system is characterized by a set of fuzzy control rules, the ordering of the rules is immaterial. This necessitates that the sentence connective also should have the properties of commutativity and associativity (see sections 111-A and 111-C in Part I and Part D in this section). In this connection, it should be noted that the operators in triangular norms and co-norms possess these properties and thus qualify as the candidates for the interpretation of the connective also. In general, we use the triangular co-norms in association with fuzzy conjunction and disjunction, and the triangular norms in association with fuzzy implication. The experimental results [52]-[54], [96], [731 and the theoretical studies [181, [851, [116], [191 relate to this issue. Kiszka et al. [52] described a preliminary investigation of the fuzzy implication functions and the sentence connective also in the context of the fuzzy model of a dc series motor. In later work, they presented additional results for fuzzy implication functions and the connective also in terms of the union and intersection operators [ S I , [MI. Our investigation leads to some preliminary conclusions. First, the connective also has a substantial influence on the quality of a fuzzy model, as we might expect. Fuzzy implication functions such as R,, R A , and R, with the connective also defined as the union operator, and R c , R,, R,,, and R,,, defined as the intersection, yield satisfactory results. These fuzzy implication functions differ in the number of mathematical operations which are needed for computer implementation. Recently, Stachowicz and Kochanska [96] studied the characteristics of 38 types of fuzzy implication along with nine different interpretations (in terms of triangular norms and co-norms) of the connective also, based on various forms of the operational curve of a series motor. Based on their results, we tabulate in Table VI a summary of the most appropriate pairs for the FLC of the fuzzy implication function and the connective also. Additional results relating to the interpretation of the connective also as the union and the intersection are reported in [73]. The investigation in question is based on a plant model with first-order delay. It is established that the fuzzy implication functions R c , R,, R h p ,R,, with the connective also as the union operator yield the best control results. Furthermore, the fuzzy implications R, and R, are not well suited for control applications even

though they yield reasonably good results in approximate reasoning. From a practical point of view, the computational aspects of an FLC require a simplification of the fuzzy control algorithm. In this perspective, Mamdanis R, and Larsens R, with the connective also as the union operator appear to be better suited for constructing fuzzy models than the other methods in FLC applications. We will have more to say about these methods at a later point. C. Compositional Operators

In a general form, a compositional operator may be expressed as the sup-star composition, where star denotes an operator-e.g., min, product, etc.-which is chosen to fit a specific application. In the literature, four kinds of compositional operators can be used in the compositional rule of inference, namely:
sup-min operation [Zadeh, 19731, sup-product operation [Kaufmann, 19751, sup-bounded-product operation [Mizumoto, 19811, sup-drastic-product operation [Mizumoto, 19811.

In FLC applications, the sup-min and sup-product compositional operators are the most frequently used. The reason is obvious, when the computational aspects of an FLC are considered. However, interesting results can be obtained if we apply the sup-product, sup-boundedproduct, and sup-drastic-product operations with different fuzzy implication functions in approximate reasoning [70], [72]. The inferred results employing these compositional operators are better than those employing the sup-min operator. Further investigation of these issues in the context of the accuracy of fuzzy models may provide interesting results. D. Inference Mechanisms
The inference mechanisms employed in an FLC are generally much simpler than those used in a typical expert system, since in an FLC the consequent of a rule is not applied to the antecedent of another. In other words, in FLC we do not employ the chaining inference mechanism, since the control actions are based on one-level forward data-driven inference (GMP). The rule base of an FLC is usually derived from expert knowledge. Typically, the rule base has the form of a

426

IEEE TKANSACTIONS O N SYSTEMS, MAN, AND (YBEKNETICS, VOL..

20, NO. 2, MAK(.H/APKIL

1990

MIMO system
R={RLIMO, RhIMO,
- 9

R,MO}

where R,,,, represents the rule: if ( x is A; a n d . . . , and y is B;) then (z, is C;; . ., zq is 0;). The antecedent forms a fuzzy set A ; x x E; in the product of R,,,, space U X . . . X V. The consequent is the union of q independent control actions. Thus the ith rule R,,,, may be represented as a fuzzy implication

process state variables and the control variable, respectively; A , , B,, and C, are linguistic values of the linguistic variables x , y, and z in the universes of discourse U, V , and W, respectively, with i = 1,2; n. The fuzzy control rule if ( x is A , and y is B , ) then (z is C,))is implemented as a fuzzy implication (relation) R, and is defined as
a ,

PR, ~P-(A,dndB,~C,)(U,V,W)
=[PA,(U)

R,,,,:

( A iX . . . X B ; )

-+

( Z,+ . . . + z q )

and

PB,(V)] + P C , ( W )

from which it follows that the rule base R may be represented as the union

={ ij [ ( A ; x . . . = { ij . . .
i= 1

XB;)-+(z,+

* * *

+zq,]}

[(A;x

XB;)-+z,)],

i=l

. U [ ( A ; x ... XB;) + z 2 ) ] ; - ,
i= I

ij [ ( A ; x . . . x B ; ) - + z q ) ] }
i=l

= { U U [(A;X
k=l i=l

* * *

xB;)+zx)])

where A , and B, is a fuzzy set A , X B, in U X K R, ( A , and B , ) C, is a fuzzy implication (relation) in U x V x W7and denotes a fuzzy implication function. The consequence C is deduced from the sup-star compositional rule of inference employing the definitions of a fuzzy implication function and the connectives and and also. In what follows, we shall consider some useful properties of the FLC inference mechanism. First, we would like to show that the sup-min operator denoted by 0 and the connective also as the union operator are commutative. Thus the fuzzy control action inferred from the complete set of fuzzy control rules is equivalent to the aggregated result derived from individual control rules. Furthermore, as will be shown later, the same properties are possessed by the sup-product operator. However, the conclusion in question does not apply when the fuzzy implication is used in its traditional logical sense [lS], [19]. More specifically, we have
-+ -+

={ ~ L I I S O 7 RG41SO

,. * . ,R B & I S O J .

Lemma I: ( A ,B )0 U
Proofi

= : ,R I
n
r=l

U= : ,(A,B ) R I .
0

In effect, the rule base R of an FLC is composed of a set with each sub-rule-base of sub-rule-bases RBL,,,, RBL,,, consisting of n fuzzy control rules with multiple process state variables and a single control variable. The general rule structure of a MIMO fuzzy system can therefore be represented as a collection of MISO fuzzy systems:

C = ( A , B ) oU R ,
=(A, B )0

U ( A , and B,
r=l

-+

C,).

R =I

w41so , , ~ 9 *

,Wl1so)
*,

The membership function pc. of the fuzzy set C is pointwise defined for all w E W by
Pc,( W ) = ( P A W > P B 4 V I ) 0 max ( P R f 4 v , w ) 1
u,v,w

where RB;,,, represents the rule: if ( x is A ; a n d . . . , and y is B i ) then ( 2 , is D;), i = 1,2; . n. Let us consider the following general form of MISO fuzzy control rules in the case of two-input/single-output fuzzy systems: input: x is A and y is B R I : if x is A , and y is B , then z is C, also R,: if x is A, and y is B, then z is C,

.PRJU,v,w),.
=

. . ,PR,,(U,V,W))
9

SUP min ( ( P A 4 U ) Pus-( v )

u,v

1 7

u,v,w

max ( P R f U ,v ,w )

.PRJ U ,v7 w )9 .
=

..

PR,( U

,v ,w))}
7

... ...
also R,: if x is A, and y is B, then z is C,
z is C

SUP max (min [ ( P A ~ ( ) , P B ~ ( v ) ) , c ~ ~ U , v , w ) ]

u , v [(,V,W

. . .,min [ ( P A , <U )
=
K,V,W

PB(

v)),

PR,{ U ?v ,w

I] }

max

{ [ ( P A ( 4 ,PB( 4)P R f U 9 V

d ]7

where x , y, and z are linguistic variables representing the

...

[( P A ( 4 P B ( V ) ) O P R , , ( U , v , w ) ] } .

LEE: FUZZY L . O G l ( ~ I N C'ONTROI. SYSTEMS: FUZ7.Y IO(il(' C'ONTK0I.I ER-PART

II

427

Therefore

C'= [ ( A ' ,B ' ) . RI] U [ ( A ' , B ' ) oR,]

. U . . . U [ ( A ' , B ' ) R,,]

0

U ( A ' ,B ' )
,=I

R,

U ( A ' , B ' ) o ( A ,and B,-+C,)

/=I

U c/f.
,=I

Lemma 2: For the fuzzy conjunctions R,, R,, R,,, and R,,, we have
( A ' ,B ' ) 0 ( A, and B,
+

C,)
As will be seen in following section, the last lemma not

= [ A ' ~ ( A , - + C , ) ] ~ [ B ' O ( ~ , - + C , )only ] simplifies the process of computation but also proif P A , X B , = PA, A PB, vides a graphic interpretation of the fuzzy inference

( A ' , B ' ) 0 ( A , and B,

=

C,)
+

mechanism in the FLC. Turning to the sup-product operator, which is denoted as -,we have the following.

[A'.( A,

C , ) ][ B ' . ( B ,
if

C,)]
PA,'PB,.

Lemma 2': ( A ' , B ' ) *U R j =

i-1

U (A',B')*Rj.
i=l

PA,XB,

Proofi
C ; = ( A ' , B ' ) o ( A ,and B,-+C,)
c~c;=(l~A,,~B,)o(llA,xB -+ , Pc,)

Lemma 2': For the fuzzy conjunctions R , , R,, R,,, and R,,, we have
( A ' ,B ' ) - ( A , and B,
=
+

C,)
+

[A'.(A,

c,)] n [ B' .( B, c,)]

if
~ A , x e= , P A , A PB,
+

= ( P A ' ,P B ' ) = ( PA', P B ' )

(min ( P A ,
min [ ( PA,
PB.)

PB,)

PC,)
+

PC, ) ? ( PB,

PC,

I]

( A ' ,B ' ) . ( A , and B,

=

C,)
+

[ A ' * (A , + C,)] . [ B'*(B,

C,)]

sup min { [ ( P A , ,
u,v

if

P A , ~ B , = PA,*PB;

'min[(PA,-+PC,)~(PB,-+)lC,)]}

supmin {min [ P ~ (PA , , ~ clC,>],

+

u,v

.min [ PB,?( PB,

Pc,I]

1
[I*B"(PB,
+

Lemma 3: If the inputs are fuzzy singletons, namely, A ' = U(), B ' = vO, then the results derived by employing Mamdani's minimum operation rule R , and Larsen's product operation rule R,, respectively, may be expressed simply as
-3

=min{[ P A " ( P A , Hence we obtain

PC,)]?

PC,)]}.

R,:
1)

ffr/\APC,(W)

R,:

ff;A Pc,(W)
ff,.*PC,(W)

R,:

2)
ff,"~PCLc,(W)

R,:

C; = [ A ' . ( A ,

C,)] n [ B'o( B,

C , ) ]. Q.E.D.

where a," = p A , ( ~ ( ) p )A S,(vo) and a;= EI.A,(U()).PB,(VO). Therefore we can assert that
n

Let us consider two special cases that follow from the preceding lemma and that play an important role in FLC applications.

R,: R,:

Pes=

U a,A Pc, ,=I

n
,=I

Lemma 3: If the inputs are fuzzy singletons, namely, A'= U(),B ' = vo, then the results dervied by employing Mamdani's minimum operation rule R , and Larsen's product operation rule R,, respectively, may be expressed simply as
R,:
%*A P C , ( W )

Pea= U f f , . P C ,

where the weighting factor (firing strength) a; is a measure of the contribution of the ith rule to the fuzzy control action. The weighting factor in question may be determined by two methods. The first uses the minimum operation in the Cartesian product, which is widely used

428

hs*[;>, ,p, l [~
. .(.................-----.--

IEEC TRANSACTIONS O N SYSTTMS, MAN. A N D (YUERNFTICS, VOL.

20,

NO.

2,

MAR(.H/APRII

1990

.....,...*...... ... .................

ys

,<

Fig. 10. Graphical interpretation of Lemma 2 under a A and R c .

Fig. 1 1 .

Graphical interpretation of Lemma 2 under a and R,,.

in FLC applications. The second employs the algebraic product in the Cartesian product, thus preserving the contribution of each input variable rather than the dominant one only. In this respect, it appears to be a reasonable choice in many FLC applications. For simplicity, assume that we have two fuzzy control rules, as follows:

and second rules may be expressed as

= P A I ( X , ) ) A PB,(YO)
a2 = P A 2 ( X O ) A

PBZ(Y0)

RI:
R,:

if x is A , and y is B , then z is C , , if x is A , and y is B , then z is C , .

Fig. 10 illustrates a graphic interpretation of Lemma 2 under R , and a:. Fig. 11 shows a graphic interpretation of Lemma 2 under R , and a : . In on-line processes, the states of a control system play an essential role in control actions. The inputs are usually measured by sensors and are crisp. In some cases it may be expedient to convert the input data into fuzzy sets. In general, however, a crisp value may be treated as a fuzzy singleton. Then the firing strengths a,and a, of the first

where k A , ( x O ) and p B f y o ) play the role of the degrees of partial match between the user-supplied data and the data in the rule base. These relations play a central role in the four types of fuzzy reasoning currently employed in FLC applications, and are described in the following. I ) Fuzzy Reasoning of the First Type -Mamdanis Minimum Operation Rule as a Fuzzy Implication Function: Fuzzy reasoning of the first type is associated with the use of Mamdanis minimum operation rule R, as a fuzzy implication function. In this mode of reasoning, the ith rule leads to the control decision
PC,W
= r

A Pc,(W)
( U ,

which implies that the membership function inferred consequence C is pointwise given by
C L A W ) = llc; v Pc;
=[I

of the

APCI(W)]V[LYZAPCZ(W)].

1.F.F: FUZZY I-OGIC' IN C'ONTROL. SYSTEMS: FUZZY 1.OGIC C'ONTROI.I.F.K--PART

!I

429

t i

XO

YO

Fig. 12. Diagrammatic representation of fuzzy reasoning 1

XO

YO

Fig. 13. Diagrammatic representation of fuzzy reasoning 2.

To obtain a deterministic control action, a defuzzification strategy is required, as will be discussed at a later point. The fuzzy reasoning process is illustrated in Fig. 12, which shows a graphic interpretation of Lemma 3 in terms of Mamdani's method R,. 2) Fuzzy Reasoning of the Second Type- Larsen's Product Operation Rule as a Fuzzy Implication Function: Fuzzy reasoning of the second type is based on the use of Larsen's product operation rule R , as a fuzzy implication function. In this case, the ith rule leads to the control decision
P ~ : ( w= ) ai*Pc,(w)-

Consequently, the membership function pc of the inferred consequence C is pointwise given by

P C ( W ) = P c ; v Pc;

process is illustrated in Fig. 13, which shows a graphic interpretation of Lemma 3 in terms of Larsen's method R,. 3) Fuzzy Reasoning of the Third Type -Tsukamoto's Method with Linguistic Terms as Monotonic Membership Functions: This method was proposed by Tsukamoto [1171. It is a simplified method based on the fuzzy reasoning of the first type in which the membership functions of fuzzy sets A , , B,, and C, are monotonic. However, in our derivation, A , and B, are not required to be monotonic but C, is. In Tsukamoto's method, the result inferred from the first rule is a , such that a I = C , ( y , ) .The result inferred from the second rule is a 2 such that a2= C , ( y 2 ) . Correspondingly, a crisp control action may be expressed as the weighted combination (Fig. 14)
aIYl+
20 =
%Y2

[ a , * P c , ( W ) ] v [%.PC,(W)].

+ a2

From C , a crisp control action can be deduced through the use of a defuzzification operator. The fuzzy reasoning

4 ) Fuzzy Reasoning of the Fourth Type-The Consequence of a Rule is a Function of Input Linguistic Variables: Fuzzy reasoning of the fourth type employs a modified

430

lEFK TRANSACTIONS ON SYSTEMS. M A N , A N U ('YBEKNETI('S, VOI..

20, NO. 2,

MARC"/APRIL

1990

'v. l&l 1 l&-r

.... .... . ...... . .... ..... . .. .

it. At present, the commonly used strategies may be described as the max criterion, the mean of maximum, and the center of area.

A. The mar criterion method The max criterion produces the point at which the possibility distribution of the control action reaches a maximum value. B. The Mean of Maximum Method (MOM) The MOM strategy generates a control action which represents the mean value of all local control actions whose membership functions reach the maximum. More specifically, in the case of a discrete universe, the control action may be expressed as
Z() =

U
X

Y
YO

min
y2

Fig. 14. Diagrammatic representation of fuzzy reasoning 3.

version of state evaluation function. In this mode of reasoning, the ith fuzzy control rule is of the form RI: i f ( x i s A , ; . . a n d y i s B,)thenz=f,(x;..,y)

]=I

c2 1
I w

where x , . . . , y, and z are linguistic variables representing process state variables and the control variable, respectively; A , , . . .,B, are linguistic values of the linguistic C. The Center of Area Method (COA) variables x , * * y in the universes of discourse U , . . ., V , The widely used COA strategy generates the center of respectively, with i = 1,2,. . -, n ; and f, is a function of the gravity of the possibility distribution of a control action. process state variables x , . . .,y defined in the input sub- In the case of a discrete universe, this method yields spaces. n For simplicity, assume that we have two fuzzy control cLZ(WI)*Wl rules as follows: /=I
e ,

where wI is the support value at which the membership function reaches the maximum value p z ( w l ) , and 1 is the number of such support values.

R I : if x is A , and y is B, then z

fl(x,y) x,y).

20 =

R,:

if x is A , and y is B, then z

c c
n
]=I

PAW])

= f2(

where n is the number of quantization levels of the output. Fig. 15 shows a graphical interpretation of various defuzzification strategies. Braae and Rutherford [5] presented a detailed analysis of various defuzzification f I( xo ,Yo 1 + u*f2( xo Yo) strategies (COA, MOM) and concluded that the COA Z" = strategy yields superior results (also see [58]). However, + ff2 This method was proposed by Takagi and Sugeno [lo31 the MOM strategy yields a better transient performance and has been applied to guide a model car smoothly along while the COA strategy yields a better steady-state pera crank-shaped track [98] and to park a car in a garage formance [94]. It should be noted that when the MOM strategy is used, the performance of an FLC is similar to [971, P91. that of a multilevel relay system [48], while the COA strategy yields results which are similar to those obtain11. DEFUZZIFICATION STRATEGIES able with a conventional PI controller [46]. An FLC based Basically, defuzzification is a mapping from a space of on the COA generally yields a lower mean square error fuzzy control actions defined over an output universe of than that based on the MOM [ l l l ] . Furthermore, the discourse into a space of nonfuzzy (crisp) control actions. MOM strategy yields a better performance than the Max It is employed because in many practical applications a criterion strategy [521. crisp control action is required. A defuzzification strategy is aimed at producing a non111. APPLICATIONS A N D RECENT DEVELOPMENTS fuzzy control action that best represents the possibility A . Applications distribution of an inferred fuzzy control action. Unfortunately, there is no systematic procedure for choosing a During the past several years, fuzzy logic has found defuzzification strategy. Zadeh 11421 first pointed out this numerous applications in fields ranging from finance to problem and made tentative suggestions for dealing with earthquake engineering [62]. In particular, fuzzy control The inferred value of the control action from the first rule is a lfl(xo,yo). The inferred value of the control action from the second rule is a2 f,<x,,, yo). Correspondingly, a crisp control action is given by
9

LEE: FlJZZY LLOCIC IN CONTROL. SYSTEMS: F1JZZY I.OGI( (ONTKOI.I.TH-IAKT

II

43 1

NL
0

Angle NM NS

ZR

PS

PM

PL

/ 1
MAX

\
MOM

Change

of
Angle

COA

Fig. 15. Diagrammatic representation of various defuzzification strate gies.

NL NM NS ZR PS PM PL

NS NM

ZR ZR PM
PS

ZR

has emerged as one of the most active and fruitful areas for research in the application of fuzzy set theory. In a:,; . ., and a:, identified by training. The training process many applications, the FLC-based systems have proved to involves a skilled operator who guides the fuzzy model car under different conditions. In this way, Sugenos car has be superior in performance to conventional systems. Notable applications of FLC include the heat exchange the capability of learning from examples. 2) FLC Hardware Systems: A higher-speed FLC hard[80], warm water process [47], activated sludge process [113], [35], traffic junction [82], cement kiln [59], [118], ware system employing fuzzy reasoning of the first type aircraft flight control [ S I , turning process [92], robot has been proposed by Yamakawa [130], 11311. It is comcontrol [1191, [941, [1061, [SI, [34], model-car parking and posed of 15 control rule boards and an action interface turning [971-[991, automobile speed control [74], [751, wa- (i.e., a defuzzifier based on the COA). It can handle fuzzy ter purification process [127], elevator control [23], auto- linguistic rules labeled as NL, N M , NS, Z R , PS, P M , PL. mobile transmission control [40], power systems and nu- The operational speed is approximately 10 mega fuzzy clear reactor control [4], [511, fuzzy memory devices [107], logical inferences per second (FLIPS). The FLC hardware system has been tested by an appli[108], [120], [1281, [1291, [133], and the fuzzy computer [132]. In this connection, it should be noted that the first cation to the stabilization of inverted pendulums mounted successful industrial application of the FLC was the ce- on a vehicle. Two pendulums with different parameters ment kiln control system developed by the Danish cement were controlled by the same set of fuzzy control rules plant manufacturer F. L. Smidth in 1979. An ingenious (Table VII). It is worthy of note that only seven fuzzy application is Sugenos fuzzy car, which has the capability control rules achieve this result. Each control rule board of learning from examples. More recently, predictive fuzzy and action interface has been integrated to a 40-pin chip. 3) Fuzzy Automatic Train Operation (ATO) Systems: control systems have been proposed and successfully applied to automatic train operation systems and automatic Hitachi Ltd. has developed a fuzzy automatic train operacontainer crane operation systems [135]-[139]. In parallel tion system (ATO) which has been in use in the Sendaiwith these developments, a great deal of progress has City subway system in Japan since July 1987. In this been made in the design of fuzzy hardware and its use in system, an object evaluation fuzzy controller predicts the performance of each candidate control command and so-called fuzzy computers [132]. selects the most likely control command based on a skilled human operators experience. B. Recent Developments More specifically, fuzzy A T 0 comprises two rule bases 1) Sugenos Fuzzy Car: One of the most interesting which evaluate two major functions of a skilled operator applications of the FLC is the fuzzy car designed by based on the criteria of safety, riding comfort, stop-gap Sugeno. Sugenos car has successfully followed a crankaccuracy, traceability of target velocity, energy consumpshaped track and parked itself in a garage [98]-[991. tion, and running time. One is constant-speed control The control policy incorporated in Sugenos car is rep(CSC), which starts a train and maintains a prescribed resented by a set of fuzzy control rules which have the speed. The other is the train automatic stop control form: (TASC), which regulates a train speed in order to stop at the target position at a station. Each rule base consists of R I : if x is A , , . . . and y is B, then twelve object-evaluation fuzzy control rules. The anz = a;, + a i x . . . + a i y tecedent of every control rule performs the evaluation of where x; . ., and y are linguistic variables representing train operation based on safety, riding comfort, stop-gap the distances and orientation in relation to the bound- accuracy, etc. The consequent determines the control aries of the track; A , ; . ., and B, are linguistic values of action to be taken based on the degree of satisfaction of x; . ., and y ; z is the value of the control variable of the each criterion. The control action is the value of the train ith control rule; and a;,; * , and a; are the parameters control notch, which is evaluated every 100 ms from the entering in the identification algorithm [ 1031, [99]. maximal evaluation of each candidate control action, and The inference mechanism of Sugenos fuzzy car is based it takes as a value a discrete number; positive value means on fuzzy reasoning of the fourth type, with the parameters power notch, negative value means break notch.

432

IELE TRANSACTIONS O N SYSTIIMS, MAN, ANI) <YHF.KNF.TI(S. V O I

. 20. N O . 2,

MAR<~ll/APRIl

1990

The Sendai-City subway system has been demonstrated to be superior in performance to the conventional PID A T 0 in riding comfort, stop gap accuracy, energy consumption, running time, and robustness [1351, [1361, [1391. 4) Fuzzy Automatic Container Crane Operation (ACO) Systems: In the application of FLC to the automatic operation of container-ship loading cranes, the principal performance criteria are safety, stop-gap accuracy, container sway, and carrying time. Fuzzy ACO involves two major operations: the trolley operation and the wire rope operation. Each operation comprises two function levels: a decision level and an activation level. Field tests of fuzzy ACO systems with real container cranes have been performed at the port of Kitakyusyu in Japan. The experimental results show that cargo handling ability of Fuzzy ACO by an unskilled operator is more than 30 containers per hour, which is comparable to the performance of a veteran operator. The tests have established that the fuzzy ACO controller has the capability of operating a crane as safely, accurately, and skillfully as a highly experienced human operator [ 1371-[ 1391. 5) Fuzzy Logic Chips and Fuzzy Computers: The first fuzzy logic chip was designed by Togai and Watanabe at AT&T Bell Laboratories in 1985 [107]. The fuzzy inference chip, which can process 16 rules in parallel, consists of four major parts: a rule-set memory, an inferenceprocessing unit, a controller, and an input-output circuitry. Recently, the rule-set memory has been implemented by a static random access memory (SRAM) to realize a capability for dynamic changes in the rule set. The inference-processing unit is based on the sup-min compositional rule of inference. Preliminary timing tests indicate that the chip can perform approximately 250000 FLIPS at 16-MHz clock. A fuzzy logic accelerator (FLA) based on this chip is currently under development [l08], [ 1201. Furthermore, in March 1989 the Microelectronics Center of North Carolina successfully completed the fabrication of the worlds fastest fuzzy logic chip, designed by Watanabe. The full-custom chip comprises 688 000 transistors and is capable of making 580000 FLIPS. In Japan, Yamakawa and Miki realized nine basic fuzzy logic functions by the standard CMOS process in current-mode circuit systems [ 1281. Later, a rudimentary concept of a fuzzy computer was proposed by Yamakawa and built by OMRON Tateishi Electric Co. Ltd [132]. The Yamakawa-OMRON computer comprises a fuzzy memory, a set of inference engines, a MAX block, a defuzzifier, and a control unit. The fuzzy memory stores linguistic fuzzy information in the form of membership functions. It has a binary RAM, a register, and a membership function generator [128]. A membership function generator (MFG) consists of a PROM, a pass transistor array, and a decoder. Every term in a term set is represented by a binary code and stored in a binary RAM. The corresponding membership functions are generated by the MFG via these binary codes. The inference engine

employs MAX and MIN operations, which are implemented by the emitter coupled fuzzy logic gates (ECFL gates) in voltage-mode circuit systems. The linguistic inputs, which are represented by analog voltages distributed on data buses, are fed into each inference engine in parallel. The results inferred from the rules are aggregated by a MAX block, which implements the function of the connective also as a union operation, yielding a consequence which is a set of analog voltages distributed on output lines. In the FLC applications, a crisp control command necessitates an auxiliary defuzzifier. In this implementation, a fuzzy computer is capable of processing fuzzy information at the very high speed of approximately 10 mega-FLIPS. It is indeed an important step not only in industrial applications but also in common-sense knowledge processing. IV. FUTURE STUDIES A N D PROBLEMS

In many of its applications, FLC is either designed by domain experts or in close collaboration with domain experts. Knowledge acquisition in FLC applications plays an important role in determining the level of performance of a fuzzy control system. However, domain experts and skilled operators do not structure their decisionmaking in any formal way. As a result, the process of transferring expert knowledge into a usable knowledge base of an FLC is time-consuming and nontrivial. Although fuzzy logic provides an effective tool for linguistic knowledge representation and Zadehs compositional rule of inference serves as a useful guideline, we are still in need of more efficient and more systematic methods for knowledge acquisition. An FLC based on the fuzzy model of a process is needed when higher accuracy and reliability are required. However, the fuzzy modeling of a process is still not well understood due to difficulties in modeling the linguistic structure of a process and obtaining operating data in industrial process control [131, [841, [llll, [1251, 11041, [loll. Classical control theory has been well developed and provides an effective tool for mathematical system analysis and design when a precise model of a system is available. In a complementary way, FLC has found many practical applications as a means of replacing a skilled human operator. For further advances, what is needed at this juncture are well-founded procedures for system design. In response to this need, many researchers are engaged in the development of a theory of fuzzy dynamic systems which extends the fundamental notions of state 161, controllability [311, and stability [771, [441, 1891, [ S I . Another direction of recent exploration is the conception and design of fuzzy systems that have the capability to learn from experience. In this area, a combination of techniques drawn from both fuzzy logic and neural network theory may provide a powerful tool for the design of systems which can emulate the remarkable human ability to learn and adapt to changes in environment.

LEE: FUZZY LOGIC IN CONTROL SYSTEMS: FUZZY LOGIC CONTROI.L.ER-PART

II

433
[241 S. Fukami. M. Mizumoto, and K. Tanaka. Some considerations of fuzzy conditional inference, Fuzzy Sets Syst., vol. 4, DD. 243-273. 1980. [25] B. R. Gaines and L. J. Kohout, The fuzzy decade: A bibliography of fuzzy systems and closely related topics, Int. J. Man. Mach. Studies, vol. 9, pp. 1-68, 1977. [26] s. Gottwald and W. Pedrycz, Problems of the design of fuzzy controllers, in Approximate Reasoning in Expert Systems, M. M. Gupta, A. Kandel, W. Bandler, and J. B. Kiszka, Ed. Amsterdam: North-Holland, 1985, pp. 393-405. [27] B. P. Graham and R. B. Newell, Fuzzy identification and control of a liquid level rig, Fuzzy Sets Syst., vol. 26, pp. 255-273, 1988. -~ Fuzzy adaptive control of a first-order process, Fuzzy Sets Syst., vol. 31, pp. 47-65, 1989. 1. Graham and P. L. Jones, Expert Sy,stems Knowledge, Uncertainty and Decision. New York: Chaman and Hall, 1988. M. M. Gupta and Y. Tsukamoto, Fuzzy logic controllersA perspective, in Proc. Joint Automatic Control Conf., San Francisco, Aug. 1980. pp. FAIO-C. 1311 M. M. Gupta, G. M. Trojan, and J. B. Kiszka, Controllability of fuzzy control systems, IEEE Trans. Syst. Man Cybern., vol. SMC-16, no 4, pp. 576-582, 1986. K. Hirota, Y. Arai, and S. Hachisu, Moving mark recognition and moving object manipulation in fuzzy controlled robot, Cont. Theory Adv. Techno[.,vol. 2, no. 3, pp. 399-418, 1986. K. Hirota and W. Pedrycz, Analysis and synthesis of fuzzy systems by the use of fuzzy sets, Fuzzy Sets Syst., vol. 10, no 1 , pp. 1-14, 1983. C. Isik,Identification and fuzzy rule-based control of a mobile robot motion, in Proc. IEEE Int. Symposium Intelligent Control, Philadelphia, PA, 1987. 0. Itoh, K. Gotoh, T. Nakayama, and S. Takamizawa, Application of fuzzy control to activated sludge process, in Proc. 2nd IFSA Congress, Tokyo, Japan, July, 1987, pp. 282-285. A. Jones, A Kaufmann, and H. J. Zimmermann, Fuzzy Sets Theory and Applications. Boston: D. Reidel Publishing Company, 1986. J. Kacprzyk and S. A. Orlovski, Optimization Models using Fuzzy Sets and Possibiliry Theory. Boston: D. Reidel Publishing Company, 1987. J. Kacprzyk and M. Fedrizzi, Combining Fuzzy Imprecision with Probabilistic Uncertainty in Decision Making. New York: Springer-Verlag, 1988. A. Kandel, Fuzzy Techniques in Pattern Recognition. New York: Wiley-Interscience, 1982. Y. Kasai and Y. Morimoto, Electronically controlled continuously variable transmission, in Proc. Int. Congress on Transportation Electronics, Dearborn, MI, 1988. A. Kaufmann, Introduction to Theory of Fuzzy Subsets. New York: Academic, 1975. A. Kaufmann and M. M. Gupta, Introduction to Fuzzy Arithmetic. New York: Van Nostrand, 1985. -, Fuzzy Mathematical Models in Engineering and Management Science. Amsterdam: North-Holland, 1988. M. Kelemen, Fuzziness and structural stability, Kybernetes, vol. IO, pp. 67-69, 1981. W. J. M. Kickert, Towards an analysis of linguistic modeling, Fuzzy Sets Syst., vol. 2, pp. 293-307, 1979. -, Further analysis and application of fuzzy logic control. Internal Rep. F/WK2/75, Queen Mary College, London, 1975. W. J. M. Kickert and H. R. Van Nauta Lemke, Application of a fuzzy controller in a warm water plant, Automat., vol. 12, no. 4, pp. 301-308, 1976. W. J. M. Kickert and E. H. Mamdani, Analysis of a fuzzy logic controller, Fuzzy Sets Syst.. vol. 1, no. 1, pp. 29-44, 1978. P. J. King and E. H. Mamdani, The application of fuzzy control systems to industrial processes, in IFAC World Congress, MIT, Boston, 1975. -, The application of fuzzy control systems to industrial processes, Automat., vol. 13, no. 3, pp. 235-242, 1977. M. Kinoshita, T. Fukuzaki, T. Satoh, and M. Miyake, An automatic operation method for control rods in BWR plants, in Proc. Specialists Meeting on In-Core Instrumentation and Reactor Core Assessment, Cadarache, France, 1988. J. B. Kiszka, M. E. Kochanska, and D. S. Sliwinska, The influence of some parameters on the accuracy of fuzzy model, in
I ,

ACKNOWLEDGMENT
1 am greatly indebted to Professor b t f i A. &de-, of the University of California-Berkeley for his encouragement of this research. The assistance of Professor Zadeh is gratefully acknowledged. The author would like to thank Professor M. Tomizuka of the University of California-Berkeley and the reviewers for their helpful cornments and suggestions.

REFERENCES
N. Baaklini and E. H. Mamdani, Prescriptive methods for deriving control policy in a fuzzy-logic controller, Electron. Lett., vol. 11, pp. 625-626, 1975. J. Baldwin and N. Guild, Modeling controllers using fuzzy relations, Kybernetes, vol. 9, pp. 223-229, 1980. J. F. Baldwin and B. W. Pilsworth, Axiomatic approach to implication for approximate reasoning with fuzzy logic, Fuzzy Sets Syst., vol 3, pp. 193-219, 1980. J. A. Bernard, Use of rule-based system for process control, IEEE Contr. Syst. Mag., vol. 8 , no. 5, pp. 3-13, 1988. M. Braae and D. A. Rutherford, Fuzzy relations in a control setting, Kybernetes, vol. 7, no. 3, pp. 185-188, 1978. -, Theoretical and linguistic aspects of the fuzzy logic controller, Automat., vol. 15, no 5, pp. 553-577, 1979. -, Section of parameters for a fuzzy logic controller, Fuzzy Sets Syst., vol. 2, no. 3, pp. 185-199, 1979. K. Ciliz, J. Fei, K. Usluel, and C. Isik, Pratical aspects of the knowledge-based control of a mobile robot motion, in Proc. 30th Midwest Symp. on Circuits and Systems, Syracuse, NY, 1987. E. Czogala, Probabilistic Sets in Decision Making and Control. Cologne: Verlag TUV Rheinland, 1984. E. Czogala, Probabilistic fuzzy controller as a generalization of the concept of fuzzy controller, Fuzzy Sets Syst., vol. 26, pp. .. 215-223, i988. . E. Czoeala and K. Hirota. Probabilistic Sets: Fuzzv and Stochastic Approach to Decision, Control and Recognition Processes. Cologne: Verlag TUV Rheinland, 1986. E. Czogala and W. Pedrycz, Some problems concerning the construction of algorithms of decision-making in fuzzy systems, Inr. J . Man. Mach. Studies, vol. 15, pp. 201-221, 1981. -, On identification in fuzzy systems and its applications in control problems, Fuzzy Sets Syst., vol. 6, no. 1, pp. 73-83, 1981. -, Control problems in fuzzy systems, Fuzzy Sets Syst., vol. 7, no. 3, pp. 257-274, 1982. -, , Fuzzy rule generation for fuzzy control, Cybem. Syst., v o l . 13, no. 3, pp. 275-293, 1982. E. Czogala and T. Rawlik, Modelling of a fuzzy controller with application to the control of biological processes, Fuzzy Sets Syst., vol. 31, pp. 13-22, 1989. B. Bharathi Devi and V. V. S. Sarma, Estimation of fuzzy memberships from histograms, Inform. Sci., vol. 35, pp. 43-59, 1985. D. Dubois and H. Prade, Fuzzy logics and the generalized modus ponens revisited, Cybern. Syst., vol. 15, pp. 3-4, 1984. -, The generalized modus ponens under sup-min composition-A theoretical study. in Approximate Reasoning in fipert Systems, M. M. Gupta, A. Kandel, W. Bandler, and J. B. Kiszka, Eds. Amsterdam: North-Holland, 1985, pp. 21 7-232. -, Unfair coins and necessity measures: Toward a possibilistic interpretation of histograms, Fuzzy Sets Syst., vol. 10, no. 1, pp. 15-20, 1985. [211 __, Possibility Theory: An Approach to Computerized Processing of Uncertainty. New York: Plenum Press, 1988. [221 J. Efstathiou, Rule-based process control using fuzzy logic, in Approximate Reasoning in Intelligent Systems, Decision and Control, E. Sanchez and L. A. Zadeh, Eds. New York: Pergamon, 1987, pp. 145-158. F. Fujitec, FLEX-8800 series elevator group control system, Fujitec Co., Ltd., Osaka, Japan, 1988.
Y

434

irrr

T R A N S A ~ ~ T I O N SON

sYs,rms, MAN.

ANI) (YHERNETICS, VOL.

20, NO. 2,

MARCH/APKII.

1990

[53] [54]

I581

I611

I631 I641

I661 1671

I701 1711 1721 I731 I741

Industrial Applications of F11zzyCvntrol. M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 187-230. -, The influence of some fuzzy implication operators on the accuracy of a fuzzy model-Part 1. Fuzzy Sets Syst., vol. 15, pp. lll-l28,1985. -, The influence of some fuzzy implication operators on the accuracy of a fuzzy model-Part 11: Fkrzy Sets Sist., vol. 15, pp. 223-240, 1985. J. B. Kiszka, M. M. Gupta, and P. N. Nikiforuk, Energetistic stability of fuzzy dynamic systems, IEEE Trans. Syst. Man Cybern., vol. SMC-15. no. 5, pp. 783-792, 1985. G. J. Klir and T. A. Folger, Fuzzy Sets, Uncertainry, and Information. Englewood Cliffs, NJ: Prentice Hall, 1988. S. V. Komolov, S. P. Makeev, and I. F. Shaknov, Optimal control of a finite automation with fuzzy constraints and a fuzzy target, Cybern., vol. 16, no. 6, pp. 805-810. 1979. L. 1. Larkin, A fuzzy logic controller for aircraft flight control, in Industrial Applications of Fuzzy Control, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 87-104. P. M. Larsen, Industrial applications of fuzzy logic control, Int. J. Man Mach. Studies, vol. 12, no. I , pp. 3-10, 1980. E. Lembessis, Dynamical learning behavior of a rule-based self-organising controller, Ph.D. thesis, Queen Mary College, Univ. of London, 1984. Y. F. Li and C. C. Lau, Development of fuzzy algorithms for servo systems, IEEE Contr. Syst. Mag., vol. 9, no. 3, pp. 65-72, 1989. J. Maiers and Y. S. Sherif, Applications of fuzzy set theory, IEEE Trans. Syst. Man Cybern., vol. SMC-15, no. I, pp. 175-189, 1985. E. H. Mamdani, Applications of fuzzy algorithms for simple dynamic plant, Proc. IEE, vol. 121, no. 12, pp. 1585-1588, 1974. E. H. Mamdani and S. Assilian, An experiment in linguistic synthesis with a fuzzy logic controller, Int. J. Man Mach. Studies, vol. 7 , no 1, pp. 1-13, 1975. E. H. Mamdani, Advances in the linguistic synthesis of fuzzy controllers, Int. J. Man Mach. Studies, vol. 8, no. 6, pp. 669-678, 1976. -, Application of fuzzy logic to approximate reasoning using linguistic synthesis, IEEE Trans. Computer. vol. C-26, no. 12, pp. 1182-1191, 1977. ~, Applications of fuzzy set theory to control systems: a survey, in Fuzzy Automata and Decision Processes, M. M. Gupta, G. N. Sardis, and B. R. Gaines, Eds. Amsterdam: NorthHolland, 1977, pp. 77-88. E. H. Mamdani and B. R. Gaines, Fuzzy Reasoning and its Applications. London: Academic, 1981. K. Matsushima and H. Sugiyama, Human operators fuzzy model in man-machine system with a nonlinear control object, in First Int. Conf. on Fuzzy Information Processing, 1984. M. Mizumoto, Note on the arithmetic rule by Zadeh for fuzzy conditional inference, Cybern. Syst., vol. 12, pp. 247-306, 1981. M. Mizumoto and H. Zimmermann, Comparison of fuzzy rea. soning methods, Fuzzy Sets and Systems, vol. 8 , p ~ 253-283, 1982. M. Mizumoto,Comparison of various fuzzy reasoning methods, in Proc. 2nd IFSA Congress, Tokyo, Japan, July 1987, pp. 2-7. -, Fuzzy controls under various approximate reasoning methods, in Proc. 2nd IFSA Congress, Tokyo, Japan, July, 1987, pp. 143-146. S. Murakami, Application of fuzzy controller to automobile speed control system, in Proc. of the IFAC Symp. on Fuzzy Information, Knowledge Representation and Decision Analysis, Marseille, France, 1983, pp. 43-48. S. Murakami and M. Maeda, Application of fuzzy controller to automobile speed control system, in Industrial Applications of Fuzzy Control, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 105-124. Y. Murayama and T. Terano, Optimizing control of diesel engine, in Industrial Applications of Fuzzy Control, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 63-72. C. V. Negoita, On the stability of fuzzy systems, in Proc. IEEE Int. Conf. on Cybernetics and Svciery, 1978, pp. 936-937. , Expert Systems and Fuzzy Systems. Menlo Park: Benjamin/Cummings Publishing Company, 1985. A. Di Nola and A. G. S. Ventre, The Mathematics of Fuzzy Systems. Gesamtherstellung: Verlag TUV Rheinland, 1986.

J. J. Ostergdad. Fuzzy logic control of a heat exchange process, in Fuzzy Automata and Decision Processes, M. M. Gupta, G. N. Saridis, and B. R. Gaines, Eds. Amsterdam: North-Holland, 1977, pp. 285-320. S. K. Pal and D. K. Dutta Majumder, Fuzzy Mathematical Approach to Pattern Recognition. New Delhi: Wiley Eastern Limited, 1986. (821 C. P. Pappis and E. H. Mamdani, A fuzzy logic controller for a traffic junction, IEEE Trans. Syst. Man Cybern., vol. SMC-7, no. 10, pp. 707-717, 1977. W. Pedrycz, An approach to the analysis of fuzzy systems, Int. J. Control, vol. 34, no. 3, pp. 403-421, 1981. ~, An identification of fuzzy relational systems, Fwzy Sets Syst., vol. 13, pp. 153-167, 1984. -, Applications of fuzzy relational equations for methods of reasoning in presence of fuzzy data, Fuzzy Sets Syst., vol. 16, pp. 163-175, 1985. H. Prade and C. V. Negoita, Fuzzy Logic in Knowledge Engineering. Gesamtherstellung: Verlag TUV Rheinland, 1986. T. J. Procyk and E. H. Mamdani, A linguistic self-organizing process controller, Automat., vol. 15, no. 1, pp. 15-30, 1979. G. P. Rao and D. A. Rutherford, Approximate reconstruction of mapping functions from linguistic description in problems of fuzzy logic applied to system control, Cybernet., vol. 12, no. 3, pp. 225-236, 1981. 1891 K. S. Ray and D. D. Majumder, Application of circle criteria for stability analysis of linear SISO and MIMO systems associated with fuzzy logic controller, IEEE Trans. Syst. Man Cybern., vol. SMC-14, no. 2, pp. 345-349, 1984. I901 D. A. Rutherford and G. C. Bloore, The implementation of fuzzy algorithms for control, in Proc. IEEE, vol. 64, no. 4, pp. 572-573, 1976. ~911 D. A. Rutherford and G. A. Carter, A Heuristic adaptive controller for a sinter plant, in Proc. 2nd IFAC Symp. on Automation in Mining, Mineral and Metal Processing, Hohannesburg, RSA, Sept. 1976. Y. Sakai, A fuzzy controller in turning process automation, in Industrial Applications of Fuzzy Control, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 139-152. E. Sanchez and L. A. Zadeh, Approximate Reasoning in Intelligent Systems, Decision and Control. New York: Pergamnon, 1988. E. M. Scharf and N. J. Mandic, The application of a fuzzy controller to the control of a multi-degree-freedom robot arm, in Industrial Applications of Fuzzy Control, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 41-62. [951 S. Shao, Fuzzy self-organizing controller and its application for dynamic processes, Fuzzy Sets Syst., vol. 26, pp. 151-164, 1988. 1961 M. S. Stachowicz and M. E. Kochanska, Fuzzy modeling of the process, in Proc. 2nd IFSA Congress, Tokyo, Japan, July 1987, pp. 86-89. M. Sugeno and K. Murakami, Fuzzy parking control of model car, in 23rd IEEE Conf. on Decision and Control, Las Vegas, 1984. M. Sugeno and M. Nishida, Fuzzy control of model car, Fuzzy Sets Syst., vol. 16, pp. 103-113, 1985. M. Sugeno and K. Murakami, An experimental study on fuzzy parking control using a model car, in Industrial Applications of Fuzzy Control, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 125-138. M. Sugeno, An introductory survey of fuzzy control, Inform. Sei., vol. 36, pp. 59-83, 1985. M. Sugeno and G. T. Kang, Structure identification of fuzzy model, Fuzzy Sets Syst., vol. 28, no. I, pp. 15-33, 1988. K. Sugiyama, Analysis and synthesis of the rule based selforganising controller, Ph.D thesis, Queen Mary College, Univ. of London, 1986. T. Takagi and M. Sugeno, Derivation of fuzzy control rules from human operators control actions, in Prvc. of the IFAC Symp. on Fuzzy Information, Knowledge Representation and Decision Analysis, Marseilles, France, July 1983, pp. 55-60. Fuzzy identification of systems and its applications to 11041 -, modeling and control, IEEE Trans. Syst. Man Cybern., vol. SMC-15, no. I, pp. 116-132, 1985. K. L. Tang and R. J. Mulholland, Comparing fuzzy logic with classical controller designs, IEEE Trans. Syst. Man Cyhern., vol. SMC-17, no. 6, pp. 1085-1087, 1987. R. Tanscheit and E. M. Scharf, Experiments with the use of a

LEE: F U Z Z Y LOGIC I N CONTROI. SYSTEMS: F U Z Z Y 1.0cilC CONTR0I.I FR-PART

II

435

[I071 [lo81 [lo91 [110] [111] [112] [113] [114] [115]

11211
[ 1221

[I281

rule-based self-organising controller for robotics applications, Fuzzy Sets Syst., vol. 26, pp. 195-214, 1988. M. Togai and H. Watanabe. Expert system on a chip: An engine for real-time approximate reasoning, IEEE Expert Syst. Mag., vol. 1, pp. 55-62, 1986. M. Togai and S. Chiu, A fuzzy accelerator and a programming environment for real-time fuzzy control, in Prvc. 2nd IFSA Congress, Tokyo, Japan, July 1987, pp. 147-151. R. M. Tong, Analysis of fuzzy control algorithms using the relation matrix, Int. J . Man Mach. Studies, vol. 8 , no. 6, pp. 679-686, 1976. -, A control engineering review of fuzzy systems, Automa., vol. 13, no. 6, pp. 559-569, 1977. -, Synthesis of fuzzy models for industrial processes, Int. Gen. Syst., vol. 4, pp. 143-162, 1978. -, The evaluation of fuzzy models derived from experimental data, Fuzzy Sets Syst., vol. 4, pp. 1-12, 1980. R. M. Tong, M. B. Beck, and A. Latten, Fuzzy control of the activated sludge wastewater treatment process, Automat., vol. 16, no. 6, pp. 695-701, 1980. R. M. Tong, A retrospective view of fuzzy control systems, Fuzzy Sets Syst., vol. 14, pp. 199-210, 1984. -, An annotated bibliography of fuzzy control, in Industrial Applications of Fuzzy Corkol, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 249-269. E. Trillas and L. Valverde, On mode and implication in approximate reasoning, in Approximate Reasoning in Expert Systems, M. M. Gupta, A. Kandel, W. Bandler, and J. B. Kiszka. Eds. Amsterdam: North-Holland, 1985, pp. 157-166. Y. Tsukamoto, An approach to fuzzy reasoning method, in Advances in Fuzzy Set Theory and Applications, M. M. Gupta, R. K. Ragade, and R. R. Yager, Eds. Amsterdam: NorthHolland, 1979. I. G. Umbers and P. J. King, An analysis of human-decision making in cement kiln control and the implications for automation, Int. J . Man Mach. Studies, vol. 12, no. 1, pp. 11-23, 1980. M. Uragami, M. Mizumoto, and K. Tananka, Fuzzy robot controls, Cybem., vol. 6, pp. 39-64, 1976. H. Watanabe and W. Dettloff, Reconfigurable fuzzy logic processor: A full custom digital VLSI, in Int. Workshop on Fuzzy System Applications, Iizuka, Japan, Aug. 1988, pp. 49-50. D. Willaeys, Optimal control of fuzzy systems, in Proc. Inr. Congress on Applied Systems Research and Cybernetics, Acapulco, Dec. 1980. -, Decision making with possibilistic dynamic systems, in Proc. 20th IEEE Conf. on Decision and Control, San Diego, Dec. 1981. D. Willaeys and N. Malvache, Use of fuzzy model for process control, in Proc. IEEE Int. Conf. on Cybernetics and Society, 1978, pp. 942-946. -, Contribution of the fuzzy sets theory to man-machine system, in Advances in Fuzzy Set Theory and Applications, M. M. Gupta, R. K. Ragade, and R. R. Yager, Eds. Amsterdam: North-Holland, 1979, pp. 481-500. C. W. Xu and Y. Zailu, Fuzzy model identification and selflearning for dynamic systems, IEEE Trans. Syst. Man Cybem., vol. SMC-17, no. 4, pp. 683-689, 1987. R. R. Yager, S. Ovchinnikov, R. M. Tong, and H. T. Nguyen, Fuzzy Sets and Applications: Selected Papers by L . A . Zadeh. New York: Wiley-Interscience, 1987. 0. Yagishita, 0. Itoh, and M. Sugeno, Application of fuzzy reasoning to the water purification process, in Industrial Applications of Fuzzy Control, M. Sugeno, Ed. Amsterdam: NorthHolland, 1985, pp. 19-40. T. Yamakawa and T. Miki, The current mode fuzzy logic integrated circuits fabricated by the standard CMOS process, IEEE Trans. Computer, vol. C-35, no. 2, pp. 161-167, 1986.

[ 1291 T. Yamakawa and K. Sasaki, Fuzzy memory device, in Prvc. 2nd IFSA Congrevs, Tokyo. Japan. July 1987, pp. 551-555. [I301 T. Yamakawa. High speed fuzzy controller hardware system. in Prvc. 2nd Fuzzy S y w m Symp.. Japan, 1986. pp. 122-130. Fuzzy controller hardware system, in Prvc. 2nd IFSA [I311 -, Congress, Tokyo. Japan, July 1987. A simple fuzzy computer hardware system employing [I321 ~, min and max operations-A challenge to 6th generation computer, in Proc. 2nd IFSA Congress, Tokyo, Japan, July 1987. [133] T. Yamakawa, Fuzzy microprocessors-Rule chip and defuzzifier chip, in Int. Workshop on Fuzzy System Applications, Iizuka, Japan, Aug. 1988, pp. 51-52. 1341 T. Yamazaki and M. Sugeno, A microprocessor based fuzzy controller for industrial purposes, in Industrial Applications of Fuzzy Control, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 231-240. 1351 S. Yasunobu, S. Miyamoto, and H. Ihara, Fuzzy control for automatic train operation system, in Prvc. 4th IFAC/IFIP/ IFORS Int. Congress on Control in Transportation Sy.stems, BadenBaden, April 1983. 1361 S. Yasunobu and S. Miyamoto, Automatic train operation by predictive fuzzy control, in Industrial Application of Fuzzy Control, M. Sugeno, Ed. Amsterdam: North-Holland, 1985, pp. 1-18. S. Yasunobu and T. Hasegawa, Evaluation of an automatic container crane operation system based on predictive fuzzy control, Control Thevty Adv. Technol., vol. 2, no. 3, pp. 419-432, 1986. ~, Predictive fuzzy control and its application for automatic container crane operation system, in Proc. 2nd IFSA Congress, Tokyo, Japan, July 1987, pp. 349-352. [139] S. Yasunobu, S. Sekino, and T. Hasegawa, Automatic train operation and automatic crane operation systems based on predictive fuzzy control, in Proc. 2nd IFSA Congress, Tokyo, Japan, July 1987, pp. 835-838. M. Yonekara, The application of fuzzy sets theory to the temperature of box annealing furnaces using simulation techniques, in Proc. 8th IFAC World Congress, Tokyo, 1981. L. A. Zadeh, Fuzzy sets, Informat. Control, vol. 8 , pp. 338-353, 1965. -, Fuzzy algorithm, Informal. Control, vol. 12, pp. 94-102, 1968. ~, Toward a theory of fuzzy systems, in Aspects of Network and System Theory, R. E. Kalman and N. DeClaris, Ed. New York: Holt, Rinehart and Winston, 1971, pp. 469-490. -, Similarity relations and fuzzy orderings, Informat. Sci., vol. 3. .. pp. 177-200, 1971. ~, A rationale for fuzzy control, Trans. ASME, J . Llynam. Syst. Measur. Control, vol. 94, pp. 3-4, 1972. -, Outline of a new approach to the analysis complex systems and decision processes, IEEE Trans. Syst. Man Cybern., vol. SMC-3, pp. 28-44, 1973. -, The concept of a linguistic variable and its application to approximate reasoning I, 11, 111, Informat. Sci., vol. 8, pp. 199-251, pp. 301-357; vol. 9, pp. 43-80, 1975. H. J. Zimmermann, Fuzzy Set Theory and Its Applications. Boston: Kluwer Academic Publishers, 1985. H. J. Zimmermann, Fuzzy Sets, Decision Making, and Grpert Systems. Boston: Kluwer Academic Publishers, 1987. C. C. Lee, Fuzzy logic in control systems: Fuzzy logic controller -Part I, IEEE Trans. Syst. Man Cybem., vol. 20, no. 2, pp. 404-418, Mar./Apr. 1990.

Chuen Chien Lee (S88) for a biography and photograph, please see page 418 of this TRANSACTIONS.