Here, x j X j, y Y, A ij F (X j ), B i F (Y ), W i [; ]. In particular, when Y is a nite set fy ; : : : ; y L g, we consider the following types of th

Size: px
Start display at page:

Download "Here, x j X j, y Y, A ij F (X j ), B i F (Y ), W i [; ]. In particular, when Y is a nite set fy ; : : : ; y L g, we consider the following types of th"

Transcription

1 A Proposal of Interpretations on Numerical Degrees of Condence for Fuzzy If{Then Rules and a Mathematical Verication of Properties under Various Reasoning Methods Tatsuya Nomura ATR Human Information Processing Research Laboratories -, Hikaridai, Seika-cho, Soraku-gun, Kyoto 69-, JAPAN Phone: , FAX: nomura@hip.atr.co.jp Abstract Some fuzzy expert systems have used fuzzy rules with numerical values which represent degrees of condence for rules. We discuss two kinds of interpretations for these numerical degrees of condence for rules, called "direct degrees " and "indirect degrees". Then, we apply Zadeh's, Baldwin's, and Tsukamoto's reasoning method to the rules under the two interpretations using general T-norms, and verify their properties. Moreover, in cases where fuzzy sets in descendant parts of rules are dened on a nite set, we present conditions for equivalence between rules with numerical degrees of condence where descendant parts are singleton form and conventional rules, under usage of {max or {sum composition for conclusions of reasoning. Introduction Some of fuzzy expert systems have recently used fuzzy rules with numerical or fuzzy values which represent degrees of condence for rules. In particular, rules with singleton types of the Then{parts and numerical degrees of condence have been used in some cases of automatic fuzzy If{Then rule extraction using neural networks or Genetic Algorithms [][][3][4]. In reasoning and composition of these rules with numerical degrees, max{min reasoning or max{product reasoning by Zadeh's direct method is basically used. However, there has been no discussions on properties in interpretations of numerical degrees of condence in reasoning by general T {norms and other methods, in particular, the relations between the conventional rules and the rules with singleton types of the Then{parts and numerical degrees of condence. In this paper, we verify the conditions for T {norms and implication functions where the conventional fuzzy If{Then rules are equivalent to the rules with singleton types of the Then{parts and numerical degrees of condence in cases where the fuzzy sets of the Then{parts are dened on a nite set, in usage of several interpretations of degrees and Zadeh's, Baldwin's, and Tsukamoto's reasoning methods. For the above purpose, we give two interpretations for numerical degrees of condence for rules (Direct and Indirect degrees) in general cases in Section. Next, we apply Zadeh's, Baldwin's, and Tsukamoto's reasoning methods to the above two interpretations and verify the relations between the conclusions of reasoning by all the methods in Section 3. Then, we derive the conditions under which the conventional rules are equivalent to the rules with singleton types of the Then{parts and numerical degrees of condence in Section 4. Direct and Indirect Degrees of Condence for Rules. Basic Descriptions In this paper, we use the following basic descriptions: F (S): The set which consists of all fuzzy sets on a set S F S : S! [; ]: The membership function of a fuzzy set F S F (S) Let X i be the i{th input set (i = ; : : : ; n), X = X X n, Y be the output set. We consider the following two types of fuzzy If{Then rules. One is the conventional type: H i : If x is A i and : : : and x n is A in ; then y is B i (i = ; : : : ; m) and another is the type made by adding numerical degrees of condence to the above H i : R i : If x is A i and : : : and x n is A in ; then y is B i with W i (i = ; : : : ; m)

2 Here, x j X j, y Y, A ij F (X j ), B i F (Y ), W i [; ]. In particular, when Y is a nite set fy ; : : : ; y L g, we consider the following types of the rules with singleton type of the Then{parts: H ki : If x is A ki and : : : and x n is A kin ; then y is y k R ki : If x is A ki and : : : and x n is A kin ; then y is y k with W ki k = ; : : : ; L; i = ; : : : ; mk ; A kij F (X j ); W ki [; ] In reasoning processes, an implication function I and a T {norm satisfying the following conditions are given : I : [; ] [; ]! [; ] T : [; ] [; ]! [; ] We describe as T (a; b) = a b. In particular, T satises the following (T){(T4): (T) a = a (8a [; ]) (T) a c; b d ) a b c d (T3) a b = b a (8a; b [; ]) (T4) a (b c) = (a b) c (8a; b; c [; ]) Here, b = (8b [; ]) is derived from (T) and (T). We do not assume any special conditions for I unless we note. Moreover, we call a fuzzy set on [; ] Fuzzy Truth Value. For F V F ([; ]), we describe F V : [; ]! [; ]. As examples for Fuzzy Truth Values, we can give the following ones: T rue (r) = r (8r [; ]) (r = ) AbsolutelyT rue (r) = otherwise Now, assume that premise fuzzy sets A i F (X i)(i = ; : : : ; n) are given. Let A = A A n F (X) be the direct product of A ; : : : ; A n by T {norm : A`(x) = A (x ) A (x ) A n (x n ) (8x = (x ; : : : ; x n ) X) In the rules H i ; R i ; H ki ; R ki, we describe as A i = A i A in and A ki = A ki A kin. We consider the following Modus Ponense: Premise : x is A Rule : H i Conclusion : y is Bi In order to derive the conclusion Bi in the above Modus Ponense, Zadeh's, Baldwin's, and Tsukamoto's reasoning methods have been proposed. In Zadeh's method (called Direct Method), the rule H i is assumed to represent a fuzzy relation A i ) B i on XY. The conclusion is given in the following way: B i (y) = sup A (x) I( Ai (x); Bi (y)) () xx In Baldwin's and Tsukamoto's methods (called Indirect Methods, or Truth Value Space Methods), rst, the fuzzy truth value for the If{part of H i "x is A i ", which is described as Ai ( F ([; ])), is determined from the premise "x is A " in the following way (called Converse Truth Qualication): Ai (a) = sup xx A i (x)=a A(x) (8a [; ]) () Next, the fuzzy truth value for the Then{part of H i "y is B i ", which is described as Bi ( F ([; ])), is determined from Ai and the implication function I. Finally, the conclusion is given in the following way (called Truth Quantication): B i (y) = Bi ( Bi (y)) (8y Y ) (3) In Baldwin's method, Bi is given in the following way: Bi (b) = sup Ai (a) I(a; b) (4) In Tsukamoto's method, the fuzzy truth value for a proposition "If x is P; then y is Q" (P F (X); Q F (Y )), which is described as P )Q F ([; ]), is dened as the following form: (r) = P )Q (r) = and Bi is given in the following way: sup (a;b)[;] I(a;b)=r (a) (b) (5) ( ; F ([; ])) Bi (b) = sup Ai (a) Ai)Bi (c) I(a;b)=c = sup Ai (a) (c) (6) I(a;b)=c In composition of a total of m conclusions B, B, : : :, B m, we consider the following two methods, the method by max B max and the method by sum B sum: B (y) = max max B i=;:::;m i (y) (y Y ) (7) B sum (y) = mx i= B i (y) (y Y ) (8). Interpretations for Numerical Degrees of Condence In this paper, we consider the following two interpretations for the degree of condence W i for the rule R i :. Direct Degrees of Condence In this interpretation, the numerical degrees are embedded in the conventional rules; that is, W i directly aects the Then{part of R i in the following way: R i : If x is A i and : : : and x n is A in ; then y is C i = W i B i

3 Here, the fuzzy set W i B i is dened as the following form: WiB i (y) = W i Bi (y) (8y Y ) (9) In this interpretation, R i is regarded as the same type as the conventional type. Thus, we can directly apply Zadeh's, Baldwin's, and Tsukamoto's methods to R i in the reasoning processes.. Indirect Degrees of Condence In this interpretation, the numerical degrees aect the implication function and the fuzzy truth values in the reasoning processes. In Zadeh's method, the conclusion of R i is given by applying the following implication function J Wi instead of I in (): J Wi (a; b) = W i I(a; b) () In Baldwin's and Tsukamoto's methods, the degrees aect the fuzzy truth value Bi in (3). In Baldwin's method, the conclusion is given by applying the above J Wi instead of I in (4). In Tsukamoto's method, the conclusion is given by applying (Ai)Bi)=Wi, dened in the following way, instead of Ai)Bi in (6). Here, on the denition of the fuzzy truth value for a proposition with a numerical degree "If x is P; then y is Q with W " (P F (X); Q F (Y ); W [; ]), (P )Q)=W F ([; ]), we propose the following three forms: W (r) = sup (a;b)[;] J W (a;b)=r (W (a) (b)) () W (r) = P )Q (r) = (r) () W 3 (r) = sup W (c) (3) c[;] W c=r 3 Properties of Several Reasoning Methods by Direct and Indirect Degrees For the premise A, the fuzzy If{Then rule H i, and the rule with the degree of condence R i in Section., we dene in the following way: ZB i : the conclusion by H i and Zadeh's method BB i : the conclusion by H i and Baldwin's method T B i : the conclusion by H i and Tsukamoto's method ZC i : the conclusion by R i and Zadeh's method in direct degrees BC i : the conclusion by R i and Baldwin's method in direct degrees T C i : the conclusion by R i and Tsukamoto's method in direct degrees IZC i : the conclusion by R i and Zadeh's method in indirect degrees IBC i : the conclusion by R i and Baldwin's method in indirect degrees I d T C i : the conclusion by R i and Tsukamoto's method in indirect degrees using the above W d (r) (d = ; ; 3) In this section, we verify the relations between the above conclusions. 3. Properties of Each Reasoning Method by Direct Degrees From the denitions of ZB i, BB i, T B i, ZC i, BC i, and T C i, their membership functions are represented in the following way: ZBi (y) = sup A (x) I( Ai (x); Bi (y)) (4) xx BBi (y) = sup Ai (a) I(a; Bi (y)) (5) T Bi (y) = sup I(a; B i (y))=c Ai (a) (c) (6) ZCi (y) = sup A (x) I( Ai (x); Ci (y)) (7) xx = sup A (x) I( Ai (x); W i Bi (y)) xx BCi (y) = sup Ai (a) I(a; Ci (y)) (8) = sup Ai (a) I(a; W i Bi (y)) T Ci (y) = sup I(a; C i (y))=c Ai (a) (c) (9) = sup I(a;W i B i (y))=c Ai (a) (c) Before the verication of the relations between the above fuzzy sets, we cite the following important lemma and theorem from [5], which show the inclusion relation between the conclusion by Zadeh's method and that by Baldwin's method: Lemma (from [5]) For any set of index M and fa z : z Mg [; ], sup (a z b) ( sup a z ) b (8b [; ]) zm zm In addition, if is left{continuous (i:e:; f b : [; ]! [; ] (f b (a) = a b) is continuous for any b [; ]), sup (a z b) = ( sup a z ) b (8b [; ]) zm zm Theorem (from [5]) ZB i BB i (i:e:; ZBi (y) BBi (y) (8r Y )). In addition, if is left{continuous, ZB i = BB i. As concrete examples of left{continuous T {norms in Lemma and Theorem, we can give "min" and the product "". In other words, the conclusion by Zadeh's method coincides with that by Baldwin's method in the conventional rules when T = min or is used.

4 Note In the above theorem, B i in the subject H i is any fuzzy set on Y. Thus, we also obtain the same results for ZC i and BC i. Moreover, we cite some lemma and theorem from [6] in the following extended form, which show the sucient conditions where the conclusion by Baldwin's method coincides with that by Tsukamoto's method in the conventional rules: Lemma (from [6]) In (5), if the following conditions are satised, then = T rue :. I(; b) I(a; b) (8a; b). I(; b) = b (8b [; ]) 3. = = T rue (Proof) From the condition, I(; r) = r. Thus, we obtain the following equation from the condition 3 and (T)(T3): (r) = sup (a b) r = r I(a;b)=r Moreover, from the conditions, and (T), I(a; b) = r ) b = I(; b) I(a; b) = r ) a b r = r Thus, (r) = r. Theorem (from [6]) If = T rue, then BB i = T B i. (Proof) T Bi (y) = sup a;c[;] I(a; B i (y))=c Ai (a) (c) = sup a;c[;] I(a; B i (y))=c Ai (a) c = sup Ai (a) I(a; Bi (y)) = BBi (y) Furthermore, we give the following lemma. Lemma 3 If the following conditions are satised, then = T rue :. = = T rue. I(a; b) = a b (8a; b [; ]) (Proof) From (T), f(a; b) [; ] : a b = rg 6= /o for 8r [; ]. Thus, we obtain the following equation from the conditions and : (r) = sup a b = r (a;b)[;] ab=r Using Lemma and 3, we obtain the following corollaries of Theorem : Corollary If the following conditions are satised, then BB i = T B i :. I(; b) I(a; b) (8a; b). I(; b) = b (8b [; ]) 3. = = T rue Corollary If the following conditions are satised, then BB i = T B i :. = = T rue. I(a; b) = a b (8a; b [; ]) As a result, the conclusion by Baldwin's method coincides with that by Tsukamoto's method in the conventional rules when = = T rue is used in (5) and, for example, I(a; b) = ab is used as the implication function. Note In the above theorem and corollaries, B i in the subject H i is any fuzzy set on Y. Thus, we also obtain the same results for ZC i and BC i. Next, we give the following theorem, which shows the relation between the conclusions from the conventional rule and the rule with the direct degree of condence by Zadeh's and Baldwin's methods: Theorem 3 If the following conditions are satised, then ZC i = W i ZB i = W i BB i = BC i :. I(a; w b) = w I(a; b) (8a; b; w [; ]) (Proof) From (T3)(T4), Lemma, the conditions,, and the equations (7)(8), we can obtain the following equations: ZCi (y) = W i sup xx n A (x) I( Ai (x); Bi (y)) = W i ZBi (y) BCi (y) = W i sup Ai (a) I(a; Bi (y)) = W i BBi (y) Furthermore, from Theorem and Note, we can obtain ZB i = BB i and ZC i = BC i. We can derive the following corollaries from Theorem, its corollaries, and Note : Corollary If the following conditions are satised, then ZC i = BC i = T C i = W i ZB i = W i BB i = W i T B i :. = T rue 3. I(a; w b) = w I(a; b) (8a; b; w [; ]) Corollary If the following conditions are satised, then ZC i = BC i = T C i = W i ZB i = W i BB i = W i T B i :. = = T rue 3. I(a; b) = a b (8a; b [; ])

5 As a result, the conclusion from the rule with the direct numerical degree is calculated from the T {norm and the conclusion from the conventional rule without the degree, under the usage of any of Zadeh's, Baldwin's, and Tsukamoto's methods, in particular, when T = min or and I(a; b) = a b. 3. Properties of Each Reasoning Method by Indirect Degrees From the denitions of IZC i, IBC i, and T C d i (d = ; ; 3), their membership functions are represented in the following way: IZCi (y) = sup A (x) W i I( Ai (x); Bi (y))() xx n IBCi (y) = sup Ai (a) W i I(a; Bi (y)) () d Id T C i (y) = sup Ai (a) W i (c) W ii(a; B i (y))=c (d = ; ; 3) () From equations ()() and Lemma, we obtain the following theorem, which shows the relation between the conclusions from the conventional rule and the rule with the indirect numerical degree by Zadeh's and Baldwin's methods: Theorem 4 If is left{continuous, then IZC i = W i ZB i = W i BB i = IBC i. Moreover, we obtain the following theorem from Theorem 3 and 4, which shows the relation between the conclusions from the conventional rule and the rules with the direct and indirect numerical degrees by Zadeh's and Baldwin's methods: Theorem 5 If the following conditions are satised, then IZC i = W i ZB i = W i BB i = IBC i = ZC i = BC i :. I(a; w b) = w I(a; b) (8a; b; w [; ]) Furthermore, we give the following theorem, which shows which shows the relation between the conclusions from the conventional rule and the rule with the indirect numerical degree by Tsukamoto's method: Theorem 6 If d W i (r) = T rue (r) = r for r s:t: fc [; ] : W i c = rg 6= /o, then I d T C i = IBC i. In addition, if is left{continuous, then I d T C i = IBC i = W i BB i = W i ZB i = IZC i. (Proof) Id T C i (y) = sup Ai (a) W d i (W i I(a; Bi (y))) = sup Ai (a) W i I(a; Bi (y)) = IBCi The latter half of the statement is trivial from Theorem 4. Here, we prepare the following lemmas in order to derive the corollaries of the above theorem. Lemma 4 If = T rue, then W 3 i (r) = T rue (r) for r s:t: fc [; ] : W i c = rg 6= /o. (Proof) W 3 i (r) = sup W i c = r c[;] W ic=r Lemma 5 If the following conditions are satised, then W i (r) = T rue (r) for r s:t: fc [; ] : W i c = rg 6= /o:. I(; b) I(a; b) (8a; b). I(; b) = b (8b [; ]) 3. = = T rue (Proof) From the condition, if W i c = r, then W i I(; c) = r. Thus, from the condition 3 and (T)(T), W i (r) = sup W i a b W i c = r W ii(a;b)=r From the conditions, and (T){(T4), W i I(a; b) = r ) W i b = W i I(; b) W i I(a; b) = r ) W i a b r = r Thus, (r) = r. Lemma 6 If the following conditions are satised, then W i (r) = T rue (r) for r s.t. fc [; ] : W i c = rg 6= /o:. = = T rue. I(a; b) = a b (8a; b [; ]) (proof) From (T), f(a; b) [; ] : W i a b = rg 6= /o (8r [; ]). Thus, from the conditions and, (r) = sup W i a b = r (a;b)[;] W iab=r From Lemma, 3, 4, 5, 6, and Theorem, 3, 6, we obtain the following corollaries: Corollary If = T rue, then I T C i = I 3 T C i = IBC i. In addition, if is left{continuous, then I T C i = I 3 T C i = IZC i = IBC i = W i ZB i = W i BB i = W i T B i. Corollary If the following conditions are satised, then I T C i = I T C i = I 3 T C i = IBC i :. I(; b) I(a; b) (8a; b). I(; b) = b (8b [; ]) 3. = = T rue In addition, if is left{continuous, then I T C i = I T C i = I 3 T C i = IBC i = IZC i = W i BB i = W i ZB i = W i T B i. Corollary 3 If the following conditions are satised, then I T C i = I T C i = I 3 T C i = IBC i :. = = T rue

6 . I(a; b) = a b (8a; b [; ]) In addition, if is left{continuous, then I T C i = I T C i = I 3 T C i = IBC i = IZC i = W i BB i = W i ZB i = W i T B i = ZC i = BC i = T C i. As a result, the conclusions from the rule with the direct and indirect numerical degree are equal to each other and are calculated from the T {norm and the conclusion from the conventional rule without the degree, under the usage of any of Zadeh's, Baldwin's, and Tsukamoto's methods, when T = min or, and I(a; b) = a b. 4 Comparison between Singleton Rules and Conventional Rules In this section, we assume that Y is a nite set fy ; y ; : : : ; y L g. For the premise A, the conventional fuzzy If{Then rules fh i g m i=, and the rules with the degree of condence fr ki g k=;:::;l in Section., we dene i=;:::;m k in the following way: ZB i : the conclusion by H i and Zadeh's method BB i : the conclusion by H i and Baldwin's method T B i : the conclusion by H i and Tsukamoto's method ZC ki : the conclusion by R ki and Zadeh's method in direct degrees BC ki : the conclusion by R ki and Baldwin's method in direct degrees T C ki : the conclusion by R ki and Tsukamoto's method in direct degrees IZC ki : the conclusion by R ki and Zadeh's method in indirect degrees IBC ki : the conclusion by R ki and Baldwin's method in indirect degrees I d T C ki : the conclusion by R ki and Tsukamoto's method in indirect degrees using the above W d (r) (d = ; ; 3) The membership functions of the above fuzzy sets are given in the following forms: ZBi (y l ) = sup A (x) I( Ai (x); Bi (y l )) (3) xx BBi (y l ) = sup Ai (a) I(a; Bi (y l )) (4) T Bi (y l ) = sup Ai (a) (c) (5) I(a; B i (y l))=c (i = ; : : : ; m; l = ; : : : ; L) ZCki (y l ) = sup A(x) I( Ai (x); W ki kl ) xx n (6) BCki (y l ) = sup Ai (a) I(a; W ki kl ) (7) T Cki (y l ) = sup Ai (a) (c) (8) I(a;W ki kl )=c IZCki (y l ) = sup xx n A(x) W ki I( Aki (x); kl ) (9) IBCki (y l ) = sup Aki (a) W ki I(a; kl ) (3) Id T C ki (y l ) = sup Aki (a) W d ki (c) (3) W ki I(a; kl )=c k = ; : : : ; L; i = ; : : : ; mk ; l = ; : : : ; L Here, kl is Kronecker's Delta. Moreover, Aki is the fuzzy truth value of the proposition "x is A ki " for the premise A, dened in (). P L Since a total of D = k= m k rules are given in the cases of both the conventional rules and the rules with degrees in Section., a total of D conclusions are obtained from their Modus Ponense. When C ki is the conclusion by H ki or R ki (k = ; : : : ; L; i = ; : : : ; m k ), these conclusions are integrated into one conclusion Cmax, or Csum ( F (Y )) based on (7) or (8). Cmax (y) = max Cki (y) (3) k=;:::;l i=;:::;m k Csum (y) = LX k= Xm k i= (8y Y ) Cki (y)! (33) Now, we consider the following special rules with singleton types of the Then{part and degrees of condence, which are derived from the conventional rules fh i g m i=: R ki : If x is A i and : : : and x n is A in ; then y is y k with W ki = Bi (y k ) (k = ; : : : ; L; i = ; : : : ; m) Figure shows the relation between the conventional rule H i and the derived rules fr ki g L. k= Moreover, we dene C max (i) and C sum (i) F (Y ) as the following forms: C (i) (y) = max max C ki (y) (34) k=;:::;l LX C (i) sum (y) = Cki (y) (35) k= (8y Y ) From (3)(33), we obtain the following equations: Cmax (y) = max (y) (36) i=;:::;m C max (i) Csum (y) = mx i= C (i) (y) (37) sum (8y Y ) We give the following theorem which represents the relations between the conventional rules and the rules with degrees derived from them: Theorem 7. If I(a; ) I(a; b) (8a; b [; ]), then C (i) max = ZBi if C pq = ZC pq (8p; q) BB i if C pq = BC pq (8p; q) (i = ; : : : ; m)

7 Ri: If x is Ai, then y is y with Wi= µ Bi( y) Therefore, max ZC ki (y l ) k=;:::;l = ZBi (y l ) max BC ki (y l ) k=;:::;l = BBi (y l ) Hi: If x is Ai, then y is Bi Bi y y y3 y4 y y y y3 y4 y Ri: If x is Ai, then y is y y y y3 y4 y R3i: If x is Ai, then y is y3 with W3i= µ Bi( y3) R4i: y y y3 y4 y If x is Ai, then y is y4 with W4i= µ Bi( y4) with Wi= µ Bi( y) y y y3 y4 y Figure : The Relation between H i and fr ki g L k= Derived from H i (case of L = 4). If at least one of the following conditions (i) and (ii) is satised, then C (i) sum = ZBi if C pq = ZC pq (8p; q) BB i if C pq = BC pq (8p; q) (i = ; : : : ; m) (i) I(a; ) = (8a > ) and Ai (x) > (8x X n ) (ii) I(a; ) = (8a) (Proof) From (6)(7) and W ki = Bi (y k ) (8k; i), ZCli (y l ) = sup xx n A (x) I( Ai (x); Bi (y l )) = ZBi (y l ) BCli (y l ) = sup Ai (a) I(a; Bi (y l )) = BBi (y l ) (Proof of ) I(r; ) I(r; Bi (y)) (8r; y). Thus, for k 6= l, ZBi (y l ) = sup xx n A(x) I( Ai (x); Bi (y l )) sup A (x) I( Ai (x); ) xx n = ZCki (y l ) BBi (y l ) = sup Ai (r) I(r; Bi (y l )) r[;] sup Ai (r) I(r; ) r[;] = BCki (y l ) (Proof of ) If (i) or (ii) is satised, the following equation is obtained for k 6= l: ZCki (y l ) = sup xx n A (x) I( Ai (x); ) = Moreover, if (i) is satised, Ai () = is derived from the denition of Ai. Thus, if (i) or (ii) is satised, the following equation is obtained for k 6= l: Therefore, BCki (y l ) = sup Ai (r) I(r; ) = r[;] LX LX k= k= ZCki (y l ) = ZBi (y l ) BCki (y l ) = BBi (y l ) We can derive the following corollaries from Theorem and Theorem Corollary : Corollary If the following conditions are satis- ed, then the conclusion from the conventional rules fh i g m i= coincides with that from the rules with degrees fr ki g k=;:::;l derived from fh i g, under the usage of the i=;:::;m reasoning by Zadeh's, Baldwin's, and Tsukamoto's methods, the {max composition, and direct degrees of condence:. I(a; ) I(a; b) (8a; b [; ]) 3. I(; b) I(a; b) (8a; b) 4. I(; b) = b (8b [; ]) 5. = = T rue Corollary If the following conditions are satis- ed, then the conclusion from the conventional rules fh i g m i= coincides with that from the rules with degrees fr ki g k=;:::;l derived from fh i g, under the usage of the i=;:::;m reasoning by Zadeh's, Baldwin's, and Tsukamoto's methods, the {sum composition, and direct degrees of condence:. (I(a; ) = (8a > ) and Ai (x) > (8x X)) or, (I(a; ) = (8a)) 3. I(; b) I(a; b) (8a; b) 4. I(; b) = b (8b [; ]) 5. = = T rue

8 Furthermore, we can derive the following corollary from Theorem 6 Corollary 3: Corollary 3 If the following conditions. = = T rue 3. I(a; b) = a b (8a; b [; ]) are satised, then (I): the conclusion from the conventional rules fh i g m i= coincides with that from the rules with degrees fr ki g k=;:::;l derived from fh i g, under the usage of i=;:::;m the reasoning by Zadeh's, Baldwin's, and Tsukamoto's methods, the {max composition, and direct and indirect degrees of condence. (II): the conclusion from the conventional rules fh i g m i= coincides with that from the rules with degrees fr ki g k=;:::;l derived from fh i g, under the usage of i=;:::;m the reasoning by Zadeh's, Baldwin's, and Tsukamoto's methods, the {sum composition, and direct and indirect degrees of condence. As a result, the capacity of the representation in the rules with the numerical degrees of condence is equal to that in the conventional rule under the usage of any of Zadeh's, Baldwin's, and Tsukamoto's methods, when T = min or, I(a; b) = a b, and the {max or {sum composition is used. Note 3 Even if the conditions in Corollary are satis- ed, the conclusion by direct degrees generally diers form that by indirect degrees, if the conditions in Corollary 3 are not satised. For example, let us consider a case of L =,n =, X = [; ], and m =. Now, we assume the following conditions for H : (8x [; )) A (x) = ; (x = ) B (y ) = p; B (y ) = q Moreover, we set = min and I(a; b) = min((?a+b); ). If the premise A ( A (x) = v 8x and v > max(p; q)) is given, then the membership functions are given in the following way: ZB (y ) = ZB (y ) = v ZC (y ) = ZC (y ) = v ZC (y ) = ZC (y ) = v Thus, ZB coincides with the conclusion of ZC and ZC by max composition. However, IZC (y ) = IZC (y ) = p IZC (y ) = IZC (y ) = q are obtained for IZC and IZC. Thus, ZB diers from the conclusion of ZC and ZC by max composition. 5 Conclusion We proposed the two interpretations on numerical degrees of condence for the fuzzy If{Then rules, direct and indirect degrees, and applied Zadeh's, Baldwin's, and Tsukamoto's methods to the reasoning process under the two interpretations using a general T {norm. Finally, we proved that the conventional rules are equivalent to the rules with singleton types of the Then{parts and numerical degrees of condence in cases where the fuzzy sets of the Then{parts are dened on a nite set, the T {norm is left{continuous (ex. min, ), the implication function is given as I(a; b) = a b, the max or sum composition is used in the integration of the conclusions, and Zadeh's, Baldwin's, or Tsukamoto's reasoning method is used, regardless of the interpretations of the degrees of condence. As future works, we consider an extension of the above results to degrees of condence represented by fuzzy truth values. Acknowledgment The author would like to thank Dr. Katsunori Shimohara at ATR Human Information Processing Research Laboratories for support, and Dr. Toru Yamaguchi at Utsunomiya University for discussions. References [] T. Tsuchiya, Y. Matsubara, and M. Nagamachi. A Learning Fuzzy Rule Parameters Using Genetic Algorithm. In Proc. 8th Fuzzy System Symposium, pages 45{48, Hiroshima, Japan, May 99. (Japanese). [] H. Ishibuchi, T. Nakashima, and T. Murata. A Fuzzy Classier System for Generating Linguistic Classi- cation Rules. In Proc. IEEE/Nagoya University WWW'95, pages {7, Nagoya, Japan, Nov 995. [3] T. Nomura and T. Miyoshi. An Adaptive Rule Extraction with the Fuzzy Self-Organizing Map and a Comparison with Other Methods. In Proc. ISUMA{ NAFIPS'95, pages 3{36, Maryland, USA, Sept 995. [4] T. Nomura and M. Miyoshi. Numerical Coding and Unfair Average Crossover in GA for Fuzzy Clustering and Their Applications for Automatic Fuzzy Rule Extraction. In Proc. IEEE/Nagoya University WWW'95, pages 3{, Nagoya, Japan, Nov 995. [5] S. Kawase and N. Yanagihara. On the Truth Space Approach and the Direct Approach in Fuzzy Reasoning. The Transaction of The Institute of Electronics, Information and Communication Engineers, J77{ A(3):53{537, Mar 994. (Japanese). [6] C. Qihao, N. Yanagihara, and S. Kawase. On Relations between Several Methods in Fuzzy Reasoning. Journal of Japan Society for Fuzzy Theory and Systems, 7(6):{8, 995. (Japanese).

A Family of Finite De Morgan Algebras

A Family of Finite De Morgan Algebras A Family of Finite De Morgan Algebras Carol L Walker Department of Mathematical Sciences New Mexico State University Las Cruces, NM 88003, USA Email: hardy@nmsuedu Elbert A Walker Department of Mathematical

More information

Fuzzy Expert Systems Lecture 6 (Fuzzy Logic )

Fuzzy Expert Systems Lecture 6 (Fuzzy Logic ) Fuzzy Expert Systems Lecture 6 (Fuzzy Logic ) Unlike Classical Logic, Fuzzy Logic is concerned, in the main, with modes of reasoning which are approximate rather than exact L. A. Zadeh Lecture 6 صفحه Summary

More information

Rule-Based Fuzzy Model

Rule-Based Fuzzy Model In rule-based fuzzy systems, the relationships between variables are represented by means of fuzzy if then rules of the following general form: Ifantecedent proposition then consequent proposition The

More information

Fuzzy Sets. Mirko Navara navara/fl/fset printe.pdf February 28, 2019

Fuzzy Sets. Mirko Navara   navara/fl/fset printe.pdf February 28, 2019 The notion of fuzzy set. Minimum about classical) sets Fuzzy ets Mirko Navara http://cmp.felk.cvut.cz/ navara/fl/fset printe.pdf February 8, 09 To aviod problems of the set theory, we restrict ourselves

More information

Fuzzy relational equation with defuzzication algorithm for the largest solution

Fuzzy relational equation with defuzzication algorithm for the largest solution Fuzzy Sets and Systems 123 (2001) 119 127 www.elsevier.com/locate/fss Fuzzy relational equation with defuzzication algorithm for the largest solution S. Kagei Department of Information and Systems, Faculty

More information

NON-MONOTONIC FUZZY REASONING J.L. CASTRO, E. TRILLAS, J.M. ZURITA. Technical Report #DECSAI-951

NON-MONOTONIC FUZZY REASONING J.L. CASTRO, E. TRILLAS, J.M. ZURITA. Technical Report #DECSAI-951 NON-MONOTONIC FUZZY REASONING J.L. CASTRO, E. TRILLAS, J.M. ZURITA Technical Report #DECSAI-951 NOVEMBER, 1995 Non-Monotonic Fuzzy Reasoning J.L Castro, E. Trillas, J.M. Zurita Depto. Ciencias de la Computacion

More information

Vector Space Basics. 1 Abstract Vector Spaces. 1. (commutativity of vector addition) u + v = v + u. 2. (associativity of vector addition)

Vector Space Basics. 1 Abstract Vector Spaces. 1. (commutativity of vector addition) u + v = v + u. 2. (associativity of vector addition) Vector Space Basics (Remark: these notes are highly formal and may be a useful reference to some students however I am also posting Ray Heitmann's notes to Canvas for students interested in a direct computational

More information

Special Classes of Fuzzy Integer Programming Models with All-Dierent Constraints

Special Classes of Fuzzy Integer Programming Models with All-Dierent Constraints Transaction E: Industrial Engineering Vol. 16, No. 1, pp. 1{10 c Sharif University of Technology, June 2009 Special Classes of Fuzzy Integer Programming Models with All-Dierent Constraints Abstract. K.

More information

Denition.9. Let a A; t 0; 1]. Then by a fuzzy point a t we mean the fuzzy subset of A given below: a t (x) = t if x = a 0 otherwise Denition.101]. A f

Denition.9. Let a A; t 0; 1]. Then by a fuzzy point a t we mean the fuzzy subset of A given below: a t (x) = t if x = a 0 otherwise Denition.101]. A f Some Properties of F -Spectrum of a Bounded Implicative BCK-Algebra A.Hasankhani Department of Mathematics, Faculty of Mathematical Sciences, Sistan and Baluchestan University, Zahedan, Iran Email:abhasan@hamoon.usb.ac.ir,

More information

On varieties generated by Weak Nilpotent Minimum t-norms

On varieties generated by Weak Nilpotent Minimum t-norms On varieties generated by Weak Nilpotent Minimum t-norms Carles Noguera IIIA-CSIC cnoguera@iiia.csic.es Francesc Esteva IIIA-CSIC esteva@iiia.csic.es Joan Gispert Universitat de Barcelona jgispertb@ub.edu

More information

Computational Intelligence Lecture 13:Fuzzy Logic

Computational Intelligence Lecture 13:Fuzzy Logic Computational Intelligence Lecture 13:Fuzzy Logic Farzaneh Abdollahi Department of Electrical Engineering Amirkabir University of Technology Fall 2011 arzaneh Abdollahi Computational Intelligence Lecture

More information

1 Outline Part I: Linear Programming (LP) Interior-Point Approach 1. Simplex Approach Comparison Part II: Semidenite Programming (SDP) Concludin

1 Outline Part I: Linear Programming (LP) Interior-Point Approach 1. Simplex Approach Comparison Part II: Semidenite Programming (SDP) Concludin Sensitivity Analysis in LP and SDP Using Interior-Point Methods E. Alper Yldrm School of Operations Research and Industrial Engineering Cornell University Ithaca, NY joint with Michael J. Todd INFORMS

More information

in a Chaotic Neural Network distributed randomness of the input in each neuron or the weight in the

in a Chaotic Neural Network distributed randomness of the input in each neuron or the weight in the Heterogeneity Enhanced Order in a Chaotic Neural Network Shin Mizutani and Katsunori Shimohara NTT Communication Science Laboratories, 2-4 Hikaridai, Seika-cho, Soraku-gun, Kyoto, 69-237 Japan shin@cslab.kecl.ntt.co.jp

More information

The Algebra of Fuzzy Truth Values

The Algebra of Fuzzy Truth Values The Algebra of Fuzzy Truth Values Carol L. Walker and Elbert A. Walker New Mexico State University Department of Mathematical Sciences Las Cruces, NM 88003, USA Abstract The purpose of this paper is to

More information

using the Hamiltonian constellations from the packing theory, i.e., the optimal sphere packing points. However, in [11] it is shown that the upper bou

using the Hamiltonian constellations from the packing theory, i.e., the optimal sphere packing points. However, in [11] it is shown that the upper bou Some 2 2 Unitary Space-Time Codes from Sphere Packing Theory with Optimal Diversity Product of Code Size 6 Haiquan Wang Genyuan Wang Xiang-Gen Xia Abstract In this correspondence, we propose some new designs

More information

[3] R.M. Colomb and C.Y.C. Chung. Very fast decision table execution of propositional

[3] R.M. Colomb and C.Y.C. Chung. Very fast decision table execution of propositional - 14 - [3] R.M. Colomb and C.Y.C. Chung. Very fast decision table execution of propositional expert systems. Proceedings of the 8th National Conference on Articial Intelligence (AAAI-9), 199, 671{676.

More information

Virtual Robust Implementation and Strategic Revealed Preference

Virtual Robust Implementation and Strategic Revealed Preference and Strategic Revealed Preference Workshop of Mathematical Economics Celebrating the 60th birthday of Aloisio Araujo IMPA Rio de Janeiro December 2006 Denitions "implementation": requires ALL equilibria

More information

(1.) For any subset P S we denote by L(P ) the abelian group of integral relations between elements of P, i.e. L(P ) := ker Z P! span Z P S S : For ea

(1.) For any subset P S we denote by L(P ) the abelian group of integral relations between elements of P, i.e. L(P ) := ker Z P! span Z P S S : For ea Torsion of dierentials on toric varieties Klaus Altmann Institut fur reine Mathematik, Humboldt-Universitat zu Berlin Ziegelstr. 13a, D-10099 Berlin, Germany. E-mail: altmann@mathematik.hu-berlin.de Abstract

More information

Approximation Bound for Fuzzy-Neural Networks with Bell Membership Function

Approximation Bound for Fuzzy-Neural Networks with Bell Membership Function Approximation Bound for Fuzzy-Neural Networks with Bell Membership Function Weimin Ma, and Guoqing Chen School of Economics and Management, Tsinghua University, Beijing, 00084, P.R. China {mawm, chengq}@em.tsinghua.edu.cn

More information

Symbolic Representations of Iterated Maps November 4, 2000 Xin-Chu Fu, Weiping Lu 2, Peter Ashwin and Jinqiao Duan 3. School of Mathematical Sciences,

Symbolic Representations of Iterated Maps November 4, 2000 Xin-Chu Fu, Weiping Lu 2, Peter Ashwin and Jinqiao Duan 3. School of Mathematical Sciences, Symbolic Representations of Iterated Maps November 4, 2000 Xin-Chu Fu, Weiping Lu 2, Peter Ashwin and Jinqiao Duan 3. School of Mathematical Sciences, Laver Building University of Exeter, Exeter EX4 4QJ,

More information

MARKOV CHAINS: STATIONARY DISTRIBUTIONS AND FUNCTIONS ON STATE SPACES. Contents

MARKOV CHAINS: STATIONARY DISTRIBUTIONS AND FUNCTIONS ON STATE SPACES. Contents MARKOV CHAINS: STATIONARY DISTRIBUTIONS AND FUNCTIONS ON STATE SPACES JAMES READY Abstract. In this paper, we rst introduce the concepts of Markov Chains and their stationary distributions. We then discuss

More information

Stochastic dominance with imprecise information

Stochastic dominance with imprecise information Stochastic dominance with imprecise information Ignacio Montes, Enrique Miranda, Susana Montes University of Oviedo, Dep. of Statistics and Operations Research. Abstract Stochastic dominance, which is

More information

Lecture 1: Introduction & Fuzzy Control I

Lecture 1: Introduction & Fuzzy Control I Lecture 1: Introduction & Fuzzy Control I Jens Kober Robert Babuška Knowledge-Based Control Systems (SC42050) Cognitive Robotics 3mE, Delft University of Technology, The Netherlands 12-02-2018 Lecture

More information

The following techniques for methods of proofs are discussed in our text: - Vacuous proof - Trivial proof

The following techniques for methods of proofs are discussed in our text: - Vacuous proof - Trivial proof Ch. 1.6 Introduction to Proofs The following techniques for methods of proofs are discussed in our text - Vacuous proof - Trivial proof - Direct proof - Indirect proof (our book calls this by contraposition)

More information

A CHARACTERIZATION OF ANCESTRAL LIMIT PROCESSES ARISING IN HAPLOID. Abstract. conditions other limit processes do appear, where multiple mergers of

A CHARACTERIZATION OF ANCESTRAL LIMIT PROCESSES ARISING IN HAPLOID. Abstract. conditions other limit processes do appear, where multiple mergers of A CHARACTERIATIO OF ACESTRAL LIMIT PROCESSES ARISIG I HAPLOID POPULATIO GEETICS MODELS M. Mohle, Johannes Gutenberg-University, Mainz and S. Sagitov 2, Chalmers University of Technology, Goteborg Abstract

More information

A Model of Consensus. in Group Decision Making. under Linguistic Assessments. F. Herrera, E. Herrera-Viedma, J.L. Verdegay

A Model of Consensus. in Group Decision Making. under Linguistic Assessments. F. Herrera, E. Herrera-Viedma, J.L. Verdegay DECSAI Deparment of Computer Science and Articial Intelligence A Model of Consensus in Group Decision Making under Linguistic Assessments F. Herrera, E. Herrera-Viedma, J.L. Verdegay Technical Report #DECSAI-94109.

More information

UNIVERSITY OF CALICUT SCHOOL OF DISTANCE EDUCATION B Sc (MATHEMATICS) I Semester Core Course. FOUNDATIONS OF MATHEMATICS (MODULE I & ii) QUESTION BANK

UNIVERSITY OF CALICUT SCHOOL OF DISTANCE EDUCATION B Sc (MATHEMATICS) I Semester Core Course. FOUNDATIONS OF MATHEMATICS (MODULE I & ii) QUESTION BANK UNIVERSITY OF CALICUT SCHOOL OF DISTANCE EDUCATION B Sc (MATHEMATICS) (2011 Admission Onwards) I Semester Core Course FOUNDATIONS OF MATHEMATICS (MODULE I & ii) QUESTION BANK 1) If A and B are two sets

More information

Fuzzy Function: Theoretical and Practical Point of View

Fuzzy Function: Theoretical and Practical Point of View EUSFLAT-LFA 2011 July 2011 Aix-les-Bains, France Fuzzy Function: Theoretical and Practical Point of View Irina Perfilieva, University of Ostrava, Inst. for Research and Applications of Fuzzy Modeling,

More information

Songklanakarin Journal of Science and Technology SJST R1 Yaqoob

Songklanakarin Journal of Science and Technology SJST R1 Yaqoob On (, qk)-intuitionistic (fuzzy ideals, fuzzy soft ideals) of subtraction algebras Journal: Songklanakarin Journal of Science Technology Manuscript ID: SJST-0-00.R Manuscript Type: Original Article Date

More information

Probabilistic ATMS. Weiru Liu, Alan Bundy. Department of Articial Intelligence, University of Edinburgh, Edinburgh EH1 1HN, UK.

Probabilistic ATMS. Weiru Liu, Alan Bundy. Department of Articial Intelligence, University of Edinburgh, Edinburgh EH1 1HN, UK. Constructing Probabilistic ATMS Using Extended Incidence Calculus Weiru Liu, Alan Bundy Department of Articial Intelligence, University of Edinburgh, Edinburgh EH1 1HN, UK. ABSTRACT This paper discusses

More information

2 THE COMPLEXITY OF TORSION-FREENESS On the other hand, the nite presentation of a group G also does not allow us to determine almost any conceivable

2 THE COMPLEXITY OF TORSION-FREENESS On the other hand, the nite presentation of a group G also does not allow us to determine almost any conceivable THE COMPUTATIONAL COMPLEXITY OF TORSION-FREENESS OF FINITELY PRESENTED GROUPS Steffen Lempp Department of Mathematics University of Wisconsin Madison, WI 53706{1388, USA Abstract. We determine the complexity

More information

Computational Intelligence Lecture 6:Fuzzy Rule Base and Fuzzy Inference Engine

Computational Intelligence Lecture 6:Fuzzy Rule Base and Fuzzy Inference Engine Computational Intelligence Lecture 6:Fuzzy Rule Base and Fuzzy Inference Engine Farzaneh Abdollahi Department of Electrical Engineering Amirkabir University of Technology Fall 200 arzaneh Abdollahi Computational

More information

4. Lecture Fuzzy Systems

4. Lecture Fuzzy Systems Soft Control (AT 3, RMA) 4. Lecture Fuzzy Systems Fuzzy Inference 4. Outline of the Lecture. Introduction to Soft Control: definition and limitations, basics of "smart" systems 2. Knowledge representation

More information

Electronic Notes in Theoretical Computer Science 18 (1998) URL: 8 pages Towards characterizing bisim

Electronic Notes in Theoretical Computer Science 18 (1998) URL:   8 pages Towards characterizing bisim Electronic Notes in Theoretical Computer Science 18 (1998) URL: http://www.elsevier.nl/locate/entcs/volume18.html 8 pages Towards characterizing bisimilarity of value-passing processes with context-free

More information

DISCRETE MATH (A LITTLE) & BASIC GROUP THEORY - PART 3/3. Contents

DISCRETE MATH (A LITTLE) & BASIC GROUP THEORY - PART 3/3. Contents DISCRETE MATH (A LITTLE) & BASIC GROUP THEORY - PART 3/3 T.K.SUBRAHMONIAN MOOTHATHU Contents 1. Cayley s Theorem 1 2. The permutation group S n 2 3. Center of a group, and centralizers 4 4. Group actions

More information

5 Set Operations, Functions, and Counting

5 Set Operations, Functions, and Counting 5 Set Operations, Functions, and Counting Let N denote the positive integers, N 0 := N {0} be the non-negative integers and Z = N 0 ( N) the positive and negative integers including 0, Q the rational numbers,

More information

CSCE 222 Discrete Structures for Computing. Review for Exam 1. Dr. Hyunyoung Lee !!!

CSCE 222 Discrete Structures for Computing. Review for Exam 1. Dr. Hyunyoung Lee !!! CSCE 222 Discrete Structures for Computing Review for Exam 1 Dr. Hyunyoung Lee 1 Topics Propositional Logic (Sections 1.1, 1.2 and 1.3) Predicate Logic (Sections 1.4 and 1.5) Rules of Inferences and Proofs

More information

Fuzzy logic Fuzzyapproximate reasoning

Fuzzy logic Fuzzyapproximate reasoning Fuzzy logic Fuzzyapproximate reasoning 3.class 3/19/2009 1 Introduction uncertain processes dynamic engineering system models fundamental of the decision making in fuzzy based real systems is the approximate

More information

Topology, Math 581, Fall 2017 last updated: November 24, Topology 1, Math 581, Fall 2017: Notes and homework Krzysztof Chris Ciesielski

Topology, Math 581, Fall 2017 last updated: November 24, Topology 1, Math 581, Fall 2017: Notes and homework Krzysztof Chris Ciesielski Topology, Math 581, Fall 2017 last updated: November 24, 2017 1 Topology 1, Math 581, Fall 2017: Notes and homework Krzysztof Chris Ciesielski Class of August 17: Course and syllabus overview. Topology

More information

The best expert versus the smartest algorithm

The best expert versus the smartest algorithm Theoretical Computer Science 34 004 361 380 www.elsevier.com/locate/tcs The best expert versus the smartest algorithm Peter Chen a, Guoli Ding b; a Department of Computer Science, Louisiana State University,

More information

On the convergence rates of genetic algorithms

On the convergence rates of genetic algorithms Theoretical Computer Science 229 (1999) 23 39 www.elsevier.com/locate/tcs On the convergence rates of genetic algorithms Jun He a;, Lishan Kang b a Department of Computer Science, Northern Jiaotong University,

More information

Logic Part II: Intuitionistic Logic and Natural Deduction

Logic Part II: Intuitionistic Logic and Natural Deduction Yesterday Remember yesterday? classical logic: reasoning about truth of formulas propositional logic: atomic sentences, composed by connectives validity and satisability can be decided by truth tables

More information

STABILITY OF INVARIANT SUBSPACES OF COMMUTING MATRICES We obtain some further results for pairs of commuting matrices. We show that a pair of commutin

STABILITY OF INVARIANT SUBSPACES OF COMMUTING MATRICES We obtain some further results for pairs of commuting matrices. We show that a pair of commutin On the stability of invariant subspaces of commuting matrices Tomaz Kosir and Bor Plestenjak September 18, 001 Abstract We study the stability of (joint) invariant subspaces of a nite set of commuting

More information

Overview. I Review of natural deduction. I Soundness and completeness. I Semantics of propositional formulas. I Soundness proof. I Completeness proof.

Overview. I Review of natural deduction. I Soundness and completeness. I Semantics of propositional formulas. I Soundness proof. I Completeness proof. Overview I Review of natural deduction. I Soundness and completeness. I Semantics of propositional formulas. I Soundness proof. I Completeness proof. Propositional formulas Grammar: ::= p j (:) j ( ^ )

More information

INTUITIONISTIC FUZZY IDEALS OF LA-SEMIGROUPS.

INTUITIONISTIC FUZZY IDEALS OF LA-SEMIGROUPS. INTUITIONISTIC FUZZY IDEALS OF LA-SEMIGROUPS. MUHAMMAD ASLAM AND SALEEM ABDULLAH Abstract. e consider the intuitionistic fuzzi cation of the concept of several ideals in LA-semigroup S, investigate some

More information

Boolean constraint satisfaction: complexity results for optimization problems with arbitrary weights

Boolean constraint satisfaction: complexity results for optimization problems with arbitrary weights Theoretical Computer Science 244 (2000) 189 203 www.elsevier.com/locate/tcs Boolean constraint satisfaction: complexity results for optimization problems with arbitrary weights Peter Jonsson Department

More information

Midterm 1. Every element of the set of functions is continuous

Midterm 1. Every element of the set of functions is continuous Econ 200 Mathematics for Economists Midterm Question.- Consider the set of functions F C(0, ) dened by { } F = f C(0, ) f(x) = ax b, a A R and b B R That is, F is a subset of the set of continuous functions

More information

Linear Algebra (part 1) : Vector Spaces (by Evan Dummit, 2017, v. 1.07) 1.1 The Formal Denition of a Vector Space

Linear Algebra (part 1) : Vector Spaces (by Evan Dummit, 2017, v. 1.07) 1.1 The Formal Denition of a Vector Space Linear Algebra (part 1) : Vector Spaces (by Evan Dummit, 2017, v. 1.07) Contents 1 Vector Spaces 1 1.1 The Formal Denition of a Vector Space.................................. 1 1.2 Subspaces...................................................

More information

The Adjoint Action of a Homotopy-associative H-space on its Loop Space.

The Adjoint Action of a Homotopy-associative H-space on its Loop Space. The Adjoint Action of a Homotopy-associative H-space on its Loop Space. Nicholas Nguyen Department of Mathematics University of Kentucky January 17th, 2014 Assumptions Unless specied, All topological spaces:

More information

Standard Koszul standardly stratied algebras

Standard Koszul standardly stratied algebras Standard Koszul standardly stratied algebras Erzsébet Lukács, András Magyar Abstract In this paper, we prove that every standard Koszul (not necessarily graded) standardly stratied algebra is also Koszul.

More information

FUZZY ASSOCIATION RULES: A TWO-SIDED APPROACH

FUZZY ASSOCIATION RULES: A TWO-SIDED APPROACH FUZZY ASSOCIATION RULES: A TWO-SIDED APPROACH M. De Cock C. Cornelis E. E. Kerre Dept. of Applied Mathematics and Computer Science Ghent University, Krijgslaan 281 (S9), B-9000 Gent, Belgium phone: +32

More information

370 Y. B. Jun generate an LI-ideal by both an LI-ideal and an element. We dene a prime LI-ideal, and give an equivalent condition for a proper LI-idea

370 Y. B. Jun generate an LI-ideal by both an LI-ideal and an element. We dene a prime LI-ideal, and give an equivalent condition for a proper LI-idea J. Korean Math. Soc. 36 (1999), No. 2, pp. 369{380 ON LI-IDEALS AND PRIME LI-IDEALS OF LATTICE IMPLICATION ALGEBRAS Young Bae Jun Abstract. As a continuation of the paper [3], in this paper we investigate

More information

Review. Propositions, propositional operators, truth tables. Logical Equivalences. Tautologies & contradictions

Review. Propositions, propositional operators, truth tables. Logical Equivalences. Tautologies & contradictions Review Propositions, propositional operators, truth tables Logical Equivalences. Tautologies & contradictions Some common logical equivalences Predicates & quantifiers Some logical equivalences involving

More information

1 Solutions to selected problems

1 Solutions to selected problems 1 Solutions to selected problems 1. Let A B R n. Show that int A int B but in general bd A bd B. Solution. Let x int A. Then there is ɛ > 0 such that B ɛ (x) A B. This shows x int B. If A = [0, 1] and

More information

0 Sets and Induction. Sets

0 Sets and Induction. Sets 0 Sets and Induction Sets A set is an unordered collection of objects, called elements or members of the set. A set is said to contain its elements. We write a A to denote that a is an element of the set

More information

Extremal Cases of the Ahlswede-Cai Inequality. A. J. Radclie and Zs. Szaniszlo. University of Nebraska-Lincoln. Department of Mathematics

Extremal Cases of the Ahlswede-Cai Inequality. A. J. Radclie and Zs. Szaniszlo. University of Nebraska-Lincoln. Department of Mathematics Extremal Cases of the Ahlswede-Cai Inequality A J Radclie and Zs Szaniszlo University of Nebraska{Lincoln Department of Mathematics 810 Oldfather Hall University of Nebraska-Lincoln Lincoln, NE 68588 1

More information

The Complexity and Approximability of Finding. Maximum Feasible Subsystems of Linear Relations. Abstract

The Complexity and Approximability of Finding. Maximum Feasible Subsystems of Linear Relations. Abstract The Complexity and Approximability of Finding Maximum Feasible Subsystems of Linear Relations Edoardo Amaldi Department of Mathematics Swiss Federal Institute of Technology CH-1015 Lausanne amaldi@dma.epfl.ch

More information

Scalings in Linear Programming: Necessary and Sufficient Conditions for Invariance

Scalings in Linear Programming: Necessary and Sufficient Conditions for Invariance BRICS RS-96-48 A. Pekeč: Scalings in Linear Programming BRICS Basic Research in Computer Science Scalings in Linear Programming: Necessary and Sufficient Conditions for Invariance Aleksandar Pekeč BRICS

More information

Fuzzy and Non-deterministic Automata Ji Mo ko January 29, 1998 Abstract An existence of an isomorphism between a category of fuzzy automata and a cate

Fuzzy and Non-deterministic Automata Ji Mo ko January 29, 1998 Abstract An existence of an isomorphism between a category of fuzzy automata and a cate University of Ostrava Institute for Research and Applications of Fuzzy Modeling Fuzzy and Non-deterministic Automata Ji Mo ko Research report No. 8 November 6, 1997 Submitted/to appear: { Supported by:

More information

On Regularity of Incline Matrices

On Regularity of Incline Matrices International Journal of Algebra, Vol. 5, 2011, no. 19, 909-924 On Regularity of Incline Matrices A. R. Meenakshi and P. Shakila Banu Department of Mathematics Karpagam University Coimbatore-641 021, India

More information

ABSTRACT SOME PROPERTIES ON FUZZY GROUPS INTROUDUCTION. preliminary definitions, and results that are required in our discussion.

ABSTRACT SOME PROPERTIES ON FUZZY GROUPS INTROUDUCTION. preliminary definitions, and results that are required in our discussion. Structures on Fuzzy Groups and L- Fuzzy Number R.Nagarajan Assistant Professor Department of Mathematics J J College of Engineering & Technology Tiruchirappalli- 620009, Tamilnadu, India A.Solairaju Associate

More information

immediately, without knowledge of the jobs that arrive later The jobs cannot be preempted, ie, once a job is scheduled (assigned to a machine), it can

immediately, without knowledge of the jobs that arrive later The jobs cannot be preempted, ie, once a job is scheduled (assigned to a machine), it can A Lower Bound for Randomized On-Line Multiprocessor Scheduling Jir Sgall Abstract We signicantly improve the previous lower bounds on the performance of randomized algorithms for on-line scheduling jobs

More information

De Nugis Groebnerialium 5: Noether, Macaulay, Jordan

De Nugis Groebnerialium 5: Noether, Macaulay, Jordan De Nugis Groebnerialium 5: Noether, Macaulay, Jordan ACA 2018 Santiago de Compostela Re-reading Macaulay References Macaulay F. S., On the Resolution of a given Modular System into Primary Systems including

More information

If L 1 and L 2 stand for the lattices of closed sets of C 1 and C 2 respectively, then the lattice of closed sets of C 1 C 2 is the direct product L 1

If L 1 and L 2 stand for the lattices of closed sets of C 1 and C 2 respectively, then the lattice of closed sets of C 1 C 2 is the direct product L 1 Direct Product Decompositions of Lattices, Closures and Relation Schemes Leonid Libkin Department of Computer and Information Science University of Pennsylvania Philadelphia, PA 19104-6389, USA Abstract

More information

A matrix over a field F is a rectangular array of elements from F. The symbol

A matrix over a field F is a rectangular array of elements from F. The symbol Chapter MATRICES Matrix arithmetic A matrix over a field F is a rectangular array of elements from F The symbol M m n (F ) denotes the collection of all m n matrices over F Matrices will usually be denoted

More information

A Survey of Lance s Weak Expectation Property, The Double Commutant Expectation Property, and The Operator System Local Lifting Property

A Survey of Lance s Weak Expectation Property, The Double Commutant Expectation Property, and The Operator System Local Lifting Property A Survey of Lance s Weak Expectation Property, The Double Commutant Expectation Property, and The Operator System Local Lifting Property Roy M. Araiza Purdue University Department of Mathematics Abstract

More information

Conditional Value-at-Risk (CVaR) Norm: Stochastic Case

Conditional Value-at-Risk (CVaR) Norm: Stochastic Case Conditional Value-at-Risk (CVaR) Norm: Stochastic Case Alexander Mafusalov, Stan Uryasev RESEARCH REPORT 03-5 Risk Management and Financial Engineering Lab Department of Industrial and Systems Engineering

More information

Soft set theoretical approach to residuated lattices. 1. Introduction. Young Bae Jun and Xiaohong Zhang

Soft set theoretical approach to residuated lattices. 1. Introduction. Young Bae Jun and Xiaohong Zhang Quasigroups and Related Systems 24 2016, 231 246 Soft set theoretical approach to residuated lattices Young Bae Jun and Xiaohong Zhang Abstract. Molodtsov's soft set theory is applied to residuated lattices.

More information

2 Z. Lonc and M. Truszczynski investigations, we use the framework of the xed-parameter complexity introduced by Downey and Fellows [Downey and Fellow

2 Z. Lonc and M. Truszczynski investigations, we use the framework of the xed-parameter complexity introduced by Downey and Fellows [Downey and Fellow Fixed-parameter complexity of semantics for logic programs ZBIGNIEW LONC Technical University of Warsaw and MIROS LAW TRUSZCZYNSKI University of Kentucky A decision problem is called parameterized if its

More information

Exercise 8.1 We have. the function is differentiable, with. f (x 0, y 0 )(u, v) = (2ax 0 + 2by 0 )u + (2bx 0 + 2cy 0 )v.

Exercise 8.1 We have. the function is differentiable, with. f (x 0, y 0 )(u, v) = (2ax 0 + 2by 0 )u + (2bx 0 + 2cy 0 )v. Exercise 8.1 We have f(x, y) f(x 0, y 0 ) = a(x 0 + x) 2 + 2b(x 0 + x)(y 0 + y) + c(y 0 + y) 2 ax 2 0 2bx 0 y 0 cy 2 0 = (2ax 0 + 2by 0 ) x + (2bx 0 + 2cy 0 ) y + (a x 2 + 2b x y + c y 2 ). By a x 2 +2b

More information

a 11 x 1 + a 12 x a 1n x n = b 1 a 21 x 1 + a 22 x a 2n x n = b 2.

a 11 x 1 + a 12 x a 1n x n = b 1 a 21 x 1 + a 22 x a 2n x n = b 2. Chapter 1 LINEAR EQUATIONS 11 Introduction to linear equations A linear equation in n unknowns x 1, x,, x n is an equation of the form a 1 x 1 + a x + + a n x n = b, where a 1, a,, a n, b are given real

More information

Learning from Examples

Learning from Examples Learning from Examples Adriano Cruz, adriano@nce.ufrj.br PPGI-UFRJ September 20 Adriano Cruz, adriano@nce.ufrj.br (PPGI-UFRJ) Learning from Examples September 20 / 40 Summary Introduction 2 Learning from

More information

ON TRANSCENDENTAL NUMBERS GENERATED BY CERTAIN INTEGER SEQUENCES

ON TRANSCENDENTAL NUMBERS GENERATED BY CERTAIN INTEGER SEQUENCES iauliai Math. Semin., 8 (16), 2013, 6369 ON TRANSCENDENTAL NUMBERS GENERATED BY CERTAIN INTEGER SEQUENCES Soichi IKEDA, Kaneaki MATSUOKA Graduate School of Mathematics, Nagoya University, Furocho, Chikusaku,

More information

55 Separable Extensions

55 Separable Extensions 55 Separable Extensions In 54, we established the foundations of Galois theory, but we have no handy criterion for determining whether a given field extension is Galois or not. Even in the quite simple

More information

agent1 c c c c c c c c c c c agent2 c agent3 c c c c c c c c c c c indirect reward an episode a detour indirect reward an episode direct reward

agent1 c c c c c c c c c c c agent2 c agent3 c c c c c c c c c c c indirect reward an episode a detour indirect reward an episode direct reward Rationality of Reward Sharing in Multi-agent Reinforcement Learning Kazuteru Miyazaki and Shigenobu Kobayashi Tokyo Institute of Technology, Department of Computational Intelligence and Systems Science,

More information

2 IM Preprint series A, No. 1/2003 If vector x(k) = (x 1 (k) x 2 (k) ::: x n (k)) denotes the time instants in which all jobs started for the k th tim

2 IM Preprint series A, No. 1/2003 If vector x(k) = (x 1 (k) x 2 (k) ::: x n (k)) denotes the time instants in which all jobs started for the k th tim Eigenvectors of interval matrices over max-plus algebra Katarna Cechlarova Institute of mathematics, Faculty ofscience, P.J. Safarik University, Jesenna 5, 041 54 Kosice, Slovakia, e-mail: cechlarova@science.upjs.sk

More information

Eects of domain characteristics on instance-based learning algorithms

Eects of domain characteristics on instance-based learning algorithms Theoretical Computer Science 298 (2003) 207 233 www.elsevier.com/locate/tcs Eects of domain characteristics on instance-based learning algorithms Seishi Okamoto, Nobuhiro Yugami Fujitsu Laboratories, 1-9-3

More information

Minimal enumerations of subsets of a nite set and the middle level problem

Minimal enumerations of subsets of a nite set and the middle level problem Discrete Applied Mathematics 114 (2001) 109 114 Minimal enumerations of subsets of a nite set and the middle level problem A.A. Evdokimov, A.L. Perezhogin 1 Sobolev Institute of Mathematics, Novosibirsk

More information

So, we can say that fuzzy proposition is a statement p which acquires a fuzzy truth value T(p) ranges from(0 to1).

So, we can say that fuzzy proposition is a statement p which acquires a fuzzy truth value T(p) ranges from(0 to1). Chapter 4 Fuzzy Proposition Main difference between classical proposition and fuzzy proposition is in the range of their truth values. The proposition value for classical proposition is either true or

More information

8 Singular Integral Operators and L p -Regularity Theory

8 Singular Integral Operators and L p -Regularity Theory 8 Singular Integral Operators and L p -Regularity Theory 8. Motivation See hand-written notes! 8.2 Mikhlin Multiplier Theorem Recall that the Fourier transformation F and the inverse Fourier transformation

More information

Vague Set Theory Applied to BM-Algebras

Vague Set Theory Applied to BM-Algebras International Journal of Algebra, Vol. 5, 2011, no. 5, 207-222 Vague Set Theory Applied to BM-Algebras A. Borumand Saeid 1 and A. Zarandi 2 1 Dept. of Math., Shahid Bahonar University of Kerman Kerman,

More information

A note on fuzzy predicate logic. Petr H jek 1. Academy of Sciences of the Czech Republic

A note on fuzzy predicate logic. Petr H jek 1. Academy of Sciences of the Czech Republic A note on fuzzy predicate logic Petr H jek 1 Institute of Computer Science, Academy of Sciences of the Czech Republic Pod vod renskou v 2, 182 07 Prague. Abstract. Recent development of mathematical fuzzy

More information

1. Brief History of Intelligent Control Systems Design Technology

1. Brief History of Intelligent Control Systems Design Technology Acknowledgments We would like to express our appreciation to Professor S.V. Ulyanov for his continuous help, value corrections and comments to the organization of this paper. We also wish to acknowledge

More information

Flags of almost ane codes

Flags of almost ane codes Flags of almost ane codes Trygve Johnsen Hugues Verdure April 0, 207 Abstract We describe a two-party wire-tap channel of type II in the framework of almost ane codes. Its cryptological performance is

More information

Average Reward Parameters

Average Reward Parameters Simulation-Based Optimization of Markov Reward Processes: Implementation Issues Peter Marbach 2 John N. Tsitsiklis 3 Abstract We consider discrete time, nite state space Markov reward processes which depend

More information

Transversals in loops. 1. Elementary properties. 1. Introduction

Transversals in loops. 1. Elementary properties. 1. Introduction Quasigroups and Related Systems 18 (2010), 43 58 Transversals in loops. 1. Elementary properties Eugene Kuznetsov Devoted to the memory of Valentin D. Belousov (1925-1988) Abstract. A new notion of a transversal

More information

Fuzzy relation equations with dual composition

Fuzzy relation equations with dual composition Fuzzy relation equations with dual composition Lenka Nosková University of Ostrava Institute for Research and Applications of Fuzzy Modeling 30. dubna 22, 701 03 Ostrava 1 Czech Republic Lenka.Noskova@osu.cz

More information

ELEMENTARY LINEAR ALGEBRA

ELEMENTARY LINEAR ALGEBRA ELEMENTARY LINEAR ALGEBRA K R MATTHEWS DEPARTMENT OF MATHEMATICS UNIVERSITY OF QUEENSLAND First Printing, 99 Chapter LINEAR EQUATIONS Introduction to linear equations A linear equation in n unknowns x,

More information

3.1 Basic properties of real numbers - continuation Inmum and supremum of a set of real numbers

3.1 Basic properties of real numbers - continuation Inmum and supremum of a set of real numbers Chapter 3 Real numbers The notion of real number was introduced in section 1.3 where the axiomatic denition of the set of all real numbers was done and some basic properties of the set of all real numbers

More information

BG/BF 1 /B/BM-algebras are congruence permutable

BG/BF 1 /B/BM-algebras are congruence permutable Mathematica Aeterna, Vol. 5, 2015, no. 2, 351-35 BG/BF 1 /B/BM-algebras are congruence permutable Andrzej Walendziak Institute of Mathematics and Physics Siedlce University, 3 Maja 54, 08-110 Siedlce,

More information

INFLUENCE OF NORMS ON THE RESULTS OF QUERY FOR INFORMATION RETRIEVAL MECHANISMS BASED ON THESAURUS. Tanja Sekulić and Petar Hotomski

INFLUENCE OF NORMS ON THE RESULTS OF QUERY FOR INFORMATION RETRIEVAL MECHANISMS BASED ON THESAURUS. Tanja Sekulić and Petar Hotomski FACTA UNIVERSITATIS (NIŠ) Ser. Math. Inform. Vol. 22, No. 2 (2007), pp. 189 199 INFLUENCE OF NORMS ON THE RESULTS OF QUERY FOR INFORMATION RETRIEVAL MECHANISMS BASED ON THESAURUS Tanja Sekulić and Petar

More information

Corollary. If v is a discrete valuation of a field K, then it is non-archimedean.

Corollary. If v is a discrete valuation of a field K, then it is non-archimedean. 24 1. Dedekind Domains and Valuations log v(n) log n does not depend on n, i.e., for all integers n>1wehavev(n) =n c with a certain constant c. This implies v(x) = x c for all rational x, andsov is equivalent

More information

Buered Probability of Exceedance: Mathematical Properties and Optimization

Buered Probability of Exceedance: Mathematical Properties and Optimization Buered Probability of Exceedance: Mathematical Properties and Optimization Alexander Mafusalov, Stan Uryasev RESEARCH REPORT 2014-1 Risk Management and Financial Engineering Lab Department of Industrial

More information

Spatial autoregression model:strong consistency

Spatial autoregression model:strong consistency Statistics & Probability Letters 65 (2003 71 77 Spatial autoregression model:strong consistency B.B. Bhattacharyya a, J.-J. Ren b, G.D. Richardson b;, J. Zhang b a Department of Statistics, North Carolina

More information

SCTT The pqr-method august 2016

SCTT The pqr-method august 2016 SCTT The pqr-method august 2016 A. Doledenok, M. Fadin, À. Menshchikov, A. Semchankau Almost all inequalities considered in our project are symmetric. Hence if plugging (a 0, b 0, c 0 ) into our inequality

More information

CO350 Linear Programming Chapter 8: Degeneracy and Finite Termination

CO350 Linear Programming Chapter 8: Degeneracy and Finite Termination CO350 Linear Programming Chapter 8: Degeneracy and Finite Termination 22th June 2005 Chapter 8: Finite Termination Recap On Monday, we established In the absence of degeneracy, the simplex method will

More information

Contents. 2.1 Vectors in R n. Linear Algebra (part 2) : Vector Spaces (by Evan Dummit, 2017, v. 2.50) 2 Vector Spaces

Contents. 2.1 Vectors in R n. Linear Algebra (part 2) : Vector Spaces (by Evan Dummit, 2017, v. 2.50) 2 Vector Spaces Linear Algebra (part 2) : Vector Spaces (by Evan Dummit, 2017, v 250) Contents 2 Vector Spaces 1 21 Vectors in R n 1 22 The Formal Denition of a Vector Space 4 23 Subspaces 6 24 Linear Combinations and

More information

Back circulant Latin squares and the inuence of a set. L F Fitina, Jennifer Seberry and Ghulam R Chaudhry. Centre for Computer Security Research

Back circulant Latin squares and the inuence of a set. L F Fitina, Jennifer Seberry and Ghulam R Chaudhry. Centre for Computer Security Research Back circulant Latin squares and the inuence of a set L F Fitina, Jennifer Seberry and Ghulam R Chaudhry Centre for Computer Security Research School of Information Technology and Computer Science University

More information

Classication of Quadratic Forms

Classication of Quadratic Forms Classication of Quadratic Forms Manuel Araújo September 14, 2011 Abstract We present the classication of quadratic forms over the rationals and then describe a partial classication of quadratic forms over

More information

Splitting a Default Theory. Hudson Turner. University of Texas at Austin.

Splitting a Default Theory. Hudson Turner. University of Texas at Austin. Splitting a Default Theory Hudson Turner Department of Computer Sciences University of Texas at Austin Austin, TX 7872-88, USA hudson@cs.utexas.edu Abstract This paper presents mathematical results that

More information