Indonesian Journal of Electrical Engineering and Computer Science V ol. 2, No . 1, Apr il 2016, pp . 187 193 DOI: 10.11591/ijeecs .v2.i1.pp187-193 187 Finding Kic king Rang e of Sepak T akra w Game: Fuzzy Logic and Dempster -Shaf er Theor y Appr oac h Andino Maseleno * , Md. Mahm ud Hasan , Muhammad Muslihudin , and T ri Susilo wati STMIK Pr ingse wu, Pr ingse wu, Lampung, Indonesia F aculty of Inf or mation T echnology , Kazakh Br itish T echnical Univ ersity , Kazakhstan * corresponding author , e-mail: andinomaseleno@y ahoo .com Abstract Sepak takr a w is pla y ed b y tw o regus , each consisting of three pla y ers . One of th e three pla y ers shall be at the bac k and he is called a T ek ong. The other tw o pla y ers shall be in front, one on the left and the other on the r ight. Ha ving v olle y kic k ed a thro w from the net b y a team mate , the ball m ust then tr a v el o v er the net to begin pla y . Dur ing the ser vice , as soon as the T ek ong kic ks the ball, all the pla y ers are allo w ed to mo v e about freely in thei r respectiv e cour ts . The no v el approach is the i nteg r ation within a Tsukamoto’ s Fuzzy reasoning and inf erences f or e vidential reasoning based on Dempster-Shaf er theor y . Sepak takr a w is a highly comple x net-barr ier kic king spor t that in v olv es dazzling displa ys of quic k refle x es , acrobat ic twists , tur ns and s w er v es of the agile human body mo v ement. Because of the humans in v olv ement in the game , the Fuzzy Logic type reasoning are the most appropr iate . The individual r ule outputs of Tsukamoto’ s Fuzzy reasoning scheme are cr isp n umbers , and theref ore , the functional relationship betw een the input v ector and the system output can be relativ ely easily identified. The result re v eals that if T ek ong is kic k f ar and front pla y er is kic k near then another regu’ s pla y er is kic k f ar , if T ek ong is kic k near and front pla y er is kic k f ar then another regu’ s pla y er is kic k near , moreo v er possibility of kic king r ange is another regu’ s pla y er is kic k f ar in kic king r ange . K e yw or ds: fuzzy logic; Dempster-Shaf er theor y; sepak takr a w; kic king r ange Cop yright c 2016 Institute of Ad v anced Engineering and Science 1. Intr oduction Mentioned in the Mala y histor ical te xt, Sejar ah Mela yu, there is a descr iption of an incident where Sultan Mansur Shahs son, Raja Muhammad, w as accidentally hit with a r attan ball b y the son of T un P er ak, in a game that w as called sepak r aga. In Thai language , it is called takr a w , meaning twine-kic k, as the ball w as made of r attan twines . The game became popul ar throughout the Southeast Asia and in the 1940s , r ules w ere estab lished and the game became officially kno wn as sepak takr a w . Sepak takr a w or kic k v olle yball is a spor t nativ e to Southeast Asia, resemb ling v olle yball, e xcept that it uses a r attan ball and only allo ws pla y ers to use their f eet and head to touch the ball. A cross betw een f ootball and v olle yball, it is a popular spor t in Thailand, Cambodia, Mala ysia, Laos , Philippines and Indonesia. The str ategies in Sepak takr a w are also v er y similar to those in v olle yball. The receiving team will attempt to pla y the takr a w ball to w ards the f ront of the net, making the best use of their three hits , to set a nd spik e the ball [1]. In pre vious w or k [2], w e used Fuzzy Logic to find kic king r ange of sepak takr a w game . The organization of the paper is as f ollo ws: sectio n 2 discusses Fuzzy Logic and Dempster-Shaf er theor y . Section 3 discusses using Fuzzy Logic and Dempster-Shaf er theor y in sepak takr a w ga me . Conclusion is presented in section 4. 2. Fuzzy Logic and Dempster -Shaf er theor y Fuzzy Logic can handle prob lems with imprecise da ta and giv e more accur ate results . Prof essor L.A. Zadeh introduced the concept of Fuzzy Logic [3]; soon after , resea rchers used this theor y f or de v eloping ne w algor ithms and decision analysis . Fuzzy sets , proposed b y Zadeh [3] Receiv ed J an uar y 21, 2016; Re vised F ebr uar y 27, 2016; Accepted March 12, 2016 Evaluation Warning : The document was created with Spire.PDF for Python.
188 ISSN: 2502-4752 as a fr ame w or k to encounter uncer tainty , v agueness and par tial tr uth, represents a deg ree of membership f or each member of the univ erse of discourse to a subset of it. Assume that the specific case of composition based on the Max-Min oper ator , then the special case of the abo v e gener al model of Fuzzy reasoning can be defined using equation 1 as B 0 ( y ) = _ x 2 A (( A 0 ( x ) ^ R ( x; y )) = max x 2 A min ( A 0 ( x ) ; R ( x; y ) (1) The Dempster-Shaf er theor y or iginated from the concept of lo w er and upper probability induced b y a m ultiv alued mapping b y Dempster [4], [5]. F ollo wing this w or k his student Glenn Shaf er [6] fur ther e xtended the theor y in his book ”A Mathematical Theor y of Evidence”, a more thorough e xplanation of belief functions . The Dempster-Shaf er theor y [6] assumes that there is a fix ed set of m utually e xclusiv e and e xhaustiv e elements called h ypotheses or propositions and symboliz ed b y the Greek letter , represented as = f h 1 ; h 2 ; :::; h n g , where h i is called a h ypothesis or proposition. A h ypothesis can be an y subset of the fr ame , in e xample , to singletons in the fr ame or to combinations of elements in the fr ame . is also called fr ame of discer nment [6]. A basic probability assignment (bpa) is represented b y a mass function m : 2 ! [0 ; 1] [6]. Whe re 2 is the po w er set of . The sum of all basic probability assignment of all subsets of the po w er set is 1 which embodies the concept that total belief has to be one [7]. Y ager and File v attempted to present a Fuzzy inf erence system based on Fuzzy Dempster-Shaf er mathematical theor y of e vidence which combining the probabilistic inf or mation in the output [8]. The basic probability assignment is a p r imitiv e of e vidence theor y . Gener ally speaking, the ter m basic probability assignment does not ref er to probability in t he classical sense . The bpa, represented b y m, defines a mapping of the po w er set to the inter v al betw een 0 and 1, where the bpa of the n ull set is 0 and the bpas of all the subsets of the po w er set is 1. In Fuzzy Logic , tw o-v alued logic often considers 0 to be f alse and 1 to be tr ue . Fuzzy Logic deals with tr uth v alues betw een 0 and 1, and these v alues are considered as the intensity or deg rees of tr uth. Dempster-Shaf er theor y pro vides a met hod to combine the pre vious measures of e vidence of diff erent sources [6]. This r ule assumes that these sources are independent. The combination: m = m 1 m 2 , also called or thogonal sum, is defined according to the Dempster’ s r ule of combi- nation [6], giv en in equation 2. It can be applied repetitiv ely when the sources are more than tw o . After the combination, a decision can be made among the diff erent h ypotheses according to the decision r ule chosen. ( m 1 m 2 )( A ) = 8 > > > > > < > > > > > : 0; A = 0 X B i \ B j = A m 1 ( B i ) m 2 ( B j ) 1 X B i \ B j 6 =0 m 1 ( B i ) m 2 ( B j ) ; A 6 = 0 (2) Where A 2 2 , B i 2 2 and B j 2 2 . T o use Dempster-Shaf er mathematical theor y of e viden ce , there m ust be the f easib le measures to deter mine basic probability assignment. 3. Using Fuzzy Logic and Dempster -Shaf er Theor y in Sepak T akra w Game When a game begins b y one ser v e , a ball can be touched b y the attac k of one time to three times . The pla y er can use a head, a bac k, legs , and an ywhere e xcept f or the ar m from the shoulder to the point of the finger . Suppose w e are giv en pla y er position and kic king r ange in the beginning of sepak takr a w game as sho wn in T ab le 1. IJEECS V ol. 2, No . 1, Apr il 2016 : 187 193 Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS ISSN: 2502-4752 189 T ab le 1. Pla y er position and kic king r ange P osition Kic king Range (m) F ar Near T ek ong 6.50 4 F ront Pla y er 7.50 2 Opponent’ s Pla y er 9.50 4.50 W e ha v e defined the realistic r ules according to the kic king r ange calculations , these r ules will become the kno wledge base of each of the prob lems considered in the sepak takr a w game . It is necessar y to sa y that the whole kno wledge does not necessar ily ha v e to be tr anslated in r ules , sometimes some of the r ules can be redundant. T ab le 2 sho ws the r ule to find kic king r ange in sepak takr a w game . T ab le 2. The Rule Rule IF AND THEN T ek ong is F ront pla y er is Opponent’ s pla y er Rule 1 Near F ar Near Rule 2 Near Near Near Rule 3 F ar F ar F ar Rule 4 F ar Near F ar [Rule 1] IF T ek ong is near AND F ront Pla y er is f ar THEN Opponent’ s Pla y er is near [Rule 2] IF T ek ong is near AND F ront Pla y er is near THEN Opponent’ s Pla y er is near [Rule 3] IF T ek ong is f ar AND F ront Pla y er is f ar THEN Opponent’ s Pla y er is f ar [Rule 4] IF T ek ong is f ar AND F ront Pla y er is near THEN Opponent’ s Pla y er is f ar Dur ing the Sepak takr a w game , both teams will mak e diff erent po w erful mo v es to kic k and spik e the ball to go to the opponent side and f all within the boundar y line of the cour t, pla y ers tr y to pla y the ball to w ard the front of the net, making the best use of their three hits to pass , set and spik e . Suppose w e are giv en 5 conditions kic king r ange in which already kno wn as sho wn in T ab le 3. T ab le 3. Kic king r ange of tek ong and front pla y er P osition Condition Condition 1 Condition 2 Condition 3 Condition 4 Condition 5 T ek ong 6 5.50 5 4.5 4.25 F ront Pla y er 2.5 3.50 4.50 5.50 6.50 Kic king r ange of tek ong is the r ange of kic king the ball from tek ong to front pla y er . Kic king r ange of front pla y er is the r ange of kic king the ball from front pla y er to opponents pla y er . Figure 1 sho ws g r aphic of kic king r ange of tek ong and front pla y er . Fuzzy Logic and Dempster-Shaf er Theor y to Find Kic king Range of Sepak T akr a w Game Evaluation Warning : The document was created with Spire.PDF for Python.
190 ISSN: 2502-4752 Figure 1. Gr aphic of kic king r ange of tek ong and front pla y er T ek ong ( T ek ong near [ x ]) = 8 > < > : 1 ; x 4 6 : 5 x 2 ; 4 < x 6 : 5 0 ; x 6 : 5 (3) ( T ek ong f ar [ x ]) = 8 > < > : 0 ; x 4 x 4 2 : 5 ; 4 < x 6 : 5 1 ; x 6 : 5 (4) Membership v alue ( T ek ong near [6]) = 6 : 5 6 2 : 5 = 0 : 20 ( T ek ong f ar [6]) = 6 4 2 : 5 = 0 : 80 Fr ont Pla y er ( F ront Pla y er near [ y ]) = 8 > < > : 1 ; y 5 : 5 7 : 5 y 5 : 5 ; 5 : 5 < y 7 : 5 0 ; y 5 : 5 (5) ( F ront Pla y er f ar [ y ]) = 8 > < > : 0 ; y 5 : 5 y 2 5 : 5 ; 5 : 5 < y 7 : 5 1 ; y 7 : 5 (6) Membership v alue ( F r ontP l ay er near [2 : 5]) = 7 : 5 2 : 5 5 : 5 = 0 : 9 ( F r ontP l ay er f ar [2 : 5]) = 2 : 5 2 5 : 5 = 0 : 09 Opponent Pla y er ( Opponent Pla y er near [ w ]) = 8 > < > : 1 ; w 4 : 5 9 : 5 w 5 ; 4 : 5 < w 9 : 5 0 ; w 9 : 5 (7) IJEECS V ol. 2, No . 1, Apr il 2016 : 187 193 Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS ISSN: 2502-4752 191 ( Opponent Pla y er f ar [ w ]) = 8 > < > : 0 ; w 4 : 5 w 4 : 5 5 ; 4 : 5 < w 9 : 5 1 ; w 9 : 5 (8) w v alue f or each r ule with min function. Figure 2 sho ws g r aphic of fuzzy membership function. Figure 2. Gr aphic of fuzzy membership function 1 = ( T ek ong near ) \ ( F ront Pla y er f ar ) ; 1 = min ( ( T ek ong near [6] \ ( F ront Pla y er f ar [2 : 5]) 1 = min (0 : 2 ; 0 : 09) ; 1 = 0 : 09 F rom the calculation abo v e , w e ha v e f our r ules . ON is opponents pla y er near in r ange , OF is opponents pla y er f ar in r ange . With Dempster-Shaf er Theor y: 1. Rule 1 m 1 f O N g = 0 : 09 ; m 1 f g = 1 0 : 09 = 0 : 91 2. Rule 2 m 2 f O N g = 0 : 2 ; m 2 f g = 1 0 : 2 = 0 : 8 The calculation of the combined m 1 and m 2 is sho wn in T ab le 4. Each cell of the tab le contains the intersection of the corresponding propositions from m 1 and m 2 along with the product of their individual belief . Fuzzy Logic and Dempster-Shaf er Theor y to Find Kic king Range of Sepak T akr a w Game Evaluation Warning : The document was created with Spire.PDF for Python.
192 ISSN: 2502-4752 T ab le 4. The first combination of the r isk of Rule 1 and Rule 2 f ON g 0.2 0.8 f ON g 0.09 f ON g 0.018 f ON g 0.07 0.91 f ON g 0.18 0.73 The first tw o bpas m 1 and m 2 are calculated to yield a ne w bpa m 3 b y a co mbination r ule as f ollo ws: m 3 f O N g = 0 : 018+0 : 18+0 : 07 1 0 = 0 : 27 ; m 3 f g = 0 : 73 1 0 = 0 : 7 3. Rule 3 m 4 f O F g = 0 : 09 ; m 4 f g = 1 0 : 09 = 0 : 91 The calculation of the combined m 3 and m 4 is sho wn in T ab le 5. Each cell of the tab le contains the intersection of the corresponding propositions from m 3 and m 4 along with the product of their individual belief . T ab le 5. The second combination of Rule 1, Rule 2 and Rule 3 f OF g 0.2 0.8 f ON g 0.27 ; 0.02 f ON g 0.25 0.73 f OF g 0.07 0.66 The second tw o bpas m 3 and m 4 are calculated to yield a ne w bpa m 5 b y a combination r ule as f ollo ws: m 5 f O F g = 0 : 07 1 0 : 02 = 0 : 07 ; m 5 f O N g = 0 : 25 1 0 : 02 = 0 : 26 ; m 5 f g = 0 : 66 1 0 : 02 = 0 : 67 4. Rule 4 m 6 f O F g = 0 : 8 ; m 6 f g = 1 0 : 8 = 0 : 2 The calculation of the combined m 5 and m 6 is sho wn in T ab le 6. Each cell of the tab le contains the intersection of the corresponding propositions from m 5 and m 6 along with the product of their individual belief . T ab le 6. The third combination of Rule 1, Rule 2, Rule 3, and Rule 4 f OF g 0.8 0.2 f OF g 0.07 f OF g 0.06 f OF g 0.01 f ON g 0.26 ; 0.21 f ON g 0.05 0.67 f OF g 0.54 0.13 The third tw o bpas m 5 and m 6 are calculated to yield a ne w bpa m 7 b y a combination r ule as f ollo ws: m 7 f O F g = 0 : 06+0 : 54+0 : 01 1 0 : 21 = 0 : 77 ; m 7 f O N g = 0 : 05 1 0 : 21 = 0 : 06 ; m 7 f O F g = 0 : 13 1 0 : 21 = 0 : 16 Finally , the final r anking of t he deg ree of belief is 0.77 > 0.06 < 0.16. The deg ree of belief is the m 7 f O F g that is equal to 0.77 which means the possibility of kic king r ange is another regu’ s pla y er is f ar in kic king r ange as sho wn in the figure 3. IJEECS V ol. 2, No . 1, Apr il 2016 : 187 193 Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS ISSN: 2502-4752 193 Figure 3. Gr aphic of fuzzy membership function 4. Conc lusion W e ha v e descr ibed a method to find kic king r ange of sepak takr a w game using Tsukamoto’ s Fuzzy reasoning and Dempster-Shaf er theor y . The v agueness present in the definition of ter ms is consistent with the inf or mation contained in the conditional r ules when obser ving some com- ple x process . Ev en though the set of linguistic v ar iab les and their meanings is compa tib le and consistent with the set of conditional r ules used, the o v er all outcome of the qualitativ e process is tr anslated into obje c t iv e and quantifiab le results . Fuzzy mathematical tools and the calculus of Fuzzy IF-THEN r ules pro vide a most useful par adigm f or the automation and implementation of an e xtensiv e body of human kno wledge heretof ore not embodied in the quantitativ e modelling process . These mathematical tools pro vide a means of shar ing, comm u nicating, and tr ansf err ing this human subjectiv e kno wledge of systems and processes . The result re v eals that if tek ong is f ar and front pla y er is near then another regu’ s pla y er is f ar , if tek ong is near and front pla y er is f ar then another regu’ s pla y er is near , moreo v er possibility of kic king r ange is another regu’ s pla y er is f ar in kic king r ange . Ref erences [1] Inter national Sepak T akr a w F eder ation. La ws of the Game Sepak T akr a w in The 24th Kings Cup Sepaktakr a w W or ld Championship 2009 Prog r am. Bangk ok, Thailand, J uly 2-7, 2009. [2] Maseleno A, Hasan MM. Finding Kic king Range of Sepak T akr a w Game: A Fuzzy Logic Approach. TELK OMNIKA, V ol.14, No .3, pp . 1-8, 2015. [3] Zadeh LA. Fuzzy Sets . Inf or mation and Control, V ol.8, pp .338-353, 1965. [4] Dempster AP . Upper and lo w er proba bilities induced b y a m ultiv alued mapping. Ann. Math. Stat. 38: 325-339, 1967. [5] Dempster AP . A Gener alization of Ba y esian inf erence . Jour nal of the Ro y al Statistical Society . 30: 205-247, 1968. [6] Shaf er G. A Mathematical Theor y of Evidence . Pr inceton Univ ersity Press , Ne w Jerse y , 1976. [7] Y ager RR. Ar it hmetic and other oper ations on Dempster-Shaf er str uctures . Inter national Jour- nal of Man-Machine Studies . 25: 357-366, 1986. [8] Y ager RR, File v P . Including Probabilistic Uncer tainty in Fuzzy Logic Controller Modeling Using Dempster-Shaf er Theor y , IEEE T r ansactions on Systems , Man, and Cyber netics . 25: 1221- 1230, 1990. Fuzzy Logic and Dempster-Shaf er Theor y to Find Kic king Range of Sepak T akr a w Game Evaluation Warning : The document was created with Spire.PDF for Python.