Indonesian J our nal of Electrical Engineering and Computer Science V ol. 11, No. 3, September 2018, pp.857 867 ISSN: 2502-4752, DOI:10.11591/ijeecs.v11.i3.pp857-867 857 Homotopy Analysis Method f or the First Order Fuzzy V olterra-Fr edholm Integr o-Differ ential Equations Ahmed A. Hamoud 1 and Kirtiwant P . Ghadle 2 1,2 Department of Mathematics, Dr . Babasaheb Ambedkar Marathw ada Uni v ersity , Aurang abad-431004 (M.S.) India 1 Department of Mathematics, T aiz Uni v ersity , T aiz, Y emen Article Inf o Article history: Recei v ed, April 28, 2018 Re vised, Jul 22, 2018 Accepted, Aug 4, 2018 K eyw ord: Homotop y analysis method Fuzzy V olterra-Fredholm inte gro-dif ferential equation Existence and uniqueness results. ABSTRA CT A fuzzy V olterra-Fredholm inte gro-dif ferential equation (FVFIDE) in a parametric case is con v erted to its related crisp case. W e use homotop y analysis method to find the approxi- mate solution of this system and hence obtain an approximation for the fuzzy solution of the FVFIDE. This paper discusses e xistence and uniqueness results and con v er gence of the proposed method. Copyright c 2018 Institute of Advanced Engineering and Science . All rights r eserved. Corresponding A uthor: Ahmed A. Hamoud Department of Mathematics, T aiz Uni v ersity , T aiz, Y emen. Email: drahmed985@yahoo.com 1. INTR ODUCTION In recent years, the topics of fuzzy inte gral equat ions which attracted increasing interest, in particular in relation to fuzzy control, ha v e been rapidly de v eloped. The concept of fuzzy numbers and arithmetic operat ions firstly introduced by Zadeh [1], and t hen by Dubois and Prade. Also, in [2] ha v e introduced the concept of inte gration of fuzzy functions. The fuzzy mapping function w as introduced by Cheng and Zadeh [1]. Moreo v er , Dubois and Prade [3] presented an elementary fuzzy calculus based on the e xtension principle. The fuzzy inte gro-dif ferential equations are a natural w ay to model uncertainty of dynamical systems. Kale v a [4] chose to define the inte gral of the fuzzy function, using the Lebesgue-type concept for inte gration. Recently , Hence v arious other methods for solving them such as using homotop y perturbation method [5], e xpansion method [6], Laplace transformation method [7], homotop y analysis method [8], dif ferential transform method [9], fix ed point theorems [10], v ariational iteration method [11]. Also, some mathematicians ha v e studied fuzzy inte gral and inte gro-dif ferential equation by numerical techniques [12]- [21] [23, 26]. As we kno w the fuzzy inte gral and inte gro-dif ferential equations are one of the important parts of the fuzzy analysis theory that play a main role in the numerical analysis. In this w ork, we will e xamine HAM to approximate the solution of the fuzzy V olterra-Fredholm inte gro- dif ferential equat ion of the second kind. The structure of this paper is or g anized as follo ws: In Section 2, we state some kno wn notations and definitions and also some theorems which are used throughout thi s paper . In Section 3, the fuzzy V olterra-Fredholm inte gro-dif ferential equation of the second kind is briefly presented. In Section 4, we con v ert a fuzzy V olterra-Fredholm inte gro-dif ferential equation of the second kind to the system of V olterra-Fredholm inte gro-dif ferential equation of the second kind in a crisp case and approximate with HAM. In Section 5, the e xistence and uniqueness results and con v er gence of the proposed method is pro v ed. In Section 6, the analytical e xample is presented illustrate the accurac y of this method. Finally , we will gi v e a report on our paper and a brief conclusion in Section 7. J ournal Homepage: http://iaescor e .com/journals/inde x.php/IJEECS Evaluation Warning : The document was created with Spire.PDF for Python.
858 ISSN: 2502-4752 2. PRELIMIN ARIES The concept of fuzzy numbers is generalized of classical real numbers and we can say that a fuzzy number is a fuzzy subset of the real li ne which has some additional properties. The concept of fuzzy number is vital for fuzzy analysis, fuzzy inte gral equations and fuzzy dif ferential equations, and a v ery helpful tool in dif ferent applications of fuzzy sets. Basic definition of fuzzy numbers is gi v en in [1, 2, 3, 27]. Definition 2..1 [2] Let us denote by R F the class of fuzzy subsets of the r eal axis u : R ! I = [0 ; 1] ; satisfying the following pr operties: u is upper semi-continuous function, u is fuzzy con ve x,i.e , u ( x + (1 ) y ) min f u ( x ) ; u ( y ) g for all x; y 2 R ; 2 [0 ; 1] , u is normal, i.e , 9 x 0 2 R for whic h u ( x 0 ) = 1 , sup u = f x 2 R j u ( x ) > 0 g is the support of the u , and its closur e cl (sup u ) is compact. Let E be the set of all fuzzy numbers on R F . The ( cut ) -le v el set of a fuzzy number u 2 E ; 0 1 , denoted by [ u ] , is defined as [ u ] = f x 2 R : u ( x ) g ; 0 < 1 ; cl (sup u ) ; = 0 : where cl (sup u = x 2 R j u ( x ) > 0) denotes the closure of the support of u . It is clear that the -le v el set of a fuzzy number is a closed and bounded interv al [ u ( ) ; u ( )] , where u ( ) denotes the left-hand end point of [ u ] , and u ( ) denotes the right-hand end point of [ u ] . Since each u 2 R can be re g arded as a fuzzy number ~ u defined by: ~ u ( t ) = 1 ; t = u 0 ; t 6 = u: An equi v alent parametric definition is also gi v en in [1] as: Definition 2..2 [2] A fuzzy number ~ u in par ametric form is a pair ( u ; u ) of functions u ( ) , u ( ) ; 0 1 , whic h satisfy the following r equir ements: u ( ) is a bounded non-decr easing left continuous function in (0 ; 1] , and right continuous at 0 , u ( ) is a bounded non-incr easing left continuous function in (0 ; 1] , and right continuous at 0 , u ( ) u ( ) ; 0 1 . A cri sp number is simply r epr esented by u ( ) = u ( ) = ; 0 1 . W e r ecall that for a < b < c whic h a; b; c 2 R ; the triangular fuzzy number u = ( a; b; c ) determined by a; b; c ar e given suc h that u ( ) = a + ( b a ) and u ( ) = c ( c b ) ar e the end points of the -le vel sets, for all 2 [0 ; 1] . The Hausdorf f distance between fuzzy numbers gi v en by D : R F R F ! R + [ f 0 g : D ( u; ) = sup 2 [0 ; 1] max fj u ( ) ( ) j ; j u ( ) ( ) jg wehre u = ( u ( ) ; u ( )) , = ( ( ) ; ( )) R is utilized in [1]. Then, it is easy to see that d is a metric in E and has the follo wing properties: D ( u + ; + ) = D ( u; ) ; 8 u; ; 2 E , D ( k u; k ) = j k j D ( u; ) ; 8 k 2 R ; u; 2 E ; D ( ! + ; + e ) D ( ! ; ) + d ( ; e ) ; 8 ! ; ; ; e 2 E , ( D ; E ) is a complete metric space. Indonesian J Elec Eng & Comp Sci V ol. 11, No. 3, September 2018: 857 -867 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 859 Definition 2..3 The function f : [ a; b ] ! R F is called a Lipsc hitz function if ther e e xists a r eal constant L 0 suc h that, for all x; t 2 [ a; b ] D ( f ( x ) ; f ( t )) L j x t j : W e r efer to L as the Lipsc hitz constant of the function f : Remark 2..1 Let u ( ) = ( u ( ) ; u ( )) ; be a fuzzy number , we tak e u c ( ) = u ( ) + u ( ) 2 ; u d ( ) = u ( ) u ( ) 2 : It is clear that u d ( ) 0 and u ( ) = u c ( ) u d ( ) and u ( ) = u c ( ) + u d ( ) ; also a fuzzy number u 2 E is said symmetric if u c ( ) is independent of for all 0 1 : Definition 2..4 Let f : R ! E be a fuzzy valued function. If for arbitr ary fixed t 0 2 R and 8 > 0 ; 9 > 0 suc h that j t t 0 j < = ) j f ( t ) f ( t 0 ) j < , f is said to be continuous. Theor em 2..2 Let f ( x ) be a fuzzy-valued function on [ a; 1 ) and it is r epr esented by ( f ( x; ) ; f ( x; )) . F or any fixed t 2 [0 ; 1] assume f ( x; ) and f ( x; ) ar e Riemann-inte gr able on [ a; b ] for e very b a , and assume ther e ar e two positive M ( ) and M ( ) suc h that R b a f ( x; ) dx M ( ) and R b a f ( x; ) dx M ( ) for e very b a . Then f ( x ) is impr oper fuzzy Riemann-inte gr able on [ a; 1 ) and the impr oper fuzzy Riemann-inte gr al is a fuzzy number . Furthermor e , we have: Z 1 a f ( x ) dx = Z 1 a f ( x; ) dx; Z 1 a f ( x; ) dx Pr oposition 2..3 [25]. If eac h of f ( x ) and g ( x ) is fuzzy-valued function and fuzzy Riemman inte gr able on = [ a; 1 ) then f ( x ) + g ( x ) is fuzzy Riemman inte gr able on . Mor eo ver , we have: Z ( f ( x ) + g ( x )) dx = Z f ( x ) dx + Z g ( x ) dx Definition 2..5 [25] The inte gr al of a fuzzy function was define by using the Riemann inte gr al concept. Let f : [ a; b ] ! E , for eac h partition P = t 0 ; t 1 ; :::; t n of [ a; b ] and for arbitr ary i 2 [ t i 1 ; t i ] ; 1 i n , suppose R p = n X i =1 f ( i )( t i t i 1 ) := max j t i t i 1 j ; 1 i n: The definite inte gr al of f ( t ) o ver [ a; b ] is Z b a f ( t ) dt = lim ! 0 R p : Pr o vided that this limit e xists in the metric d . If the fuzzy function f ( t ) is continuous in the metric d , its definit e inte gr al e xists, and also Z b a f ( t; r ) dt = Z b a f ( t; r ) dt; Z b a f ( t; r ) dt = Z b a f ( t; r ) dt: More details about the properties of the fuzzy inte gral are gi v en in [2, 26, 25]. Theor em 2..4 [28] (Banac h contr action principle). Let ( X ; d ) be a complete metric space , then eac h contr act ion mapping T : X ! X has a unique fixed point x of T in X i.e . T x = x: Theor em 2..5 [24] (Sc hauder’ s fixed point theor em). Let X be a Banac h space and let A a con ve x, closed subset of X . If T : A ! A be the map suc h that the set f T u : u 2 A g is r elatively compact in X (or T is continuous and completely continouous). Then T has at least one fixed point u 2 A : T u = u : Homotopy Analysis Method for the F ir st Or der Fuzzy V olterr a-F r edholm... (Ahmed A. Hamoud) Evaluation Warning : The document was created with Spire.PDF for Python.
860 ISSN: 2502-4752 3. FUZZY V OL TERRA-FREDHOLM INTEGR O-DIFFERENTIAL EQ U A TION In this section we consider the fuzzy V olterra-Fredholm inte gro-dif ferential equation ~ u 0 ( x ) = ~ f ( x ) + Z x a k 1 ( x; t ) F 1 ( ~ u ( t )) dt + Z a k 2 ( x; t ) F 2 ( ~ u ( t )) dt; (1) with initial condition ~ u 0 (0) = ~ u (0) ; (2) where ; 2 R , f ( x ) ; k 1 ; k 2 and F 1 ( ~ u ( t )) are analytical functions k 1 ; k 2 : C ([0 ; ] 2 ) ! R + , that ha v e suitable deri v ati v es on an interv al 0 t x and ~ u ( x ) is unkno wn function. The solution is e xpressed in the form: ~ u ( x ) = 1 X i =0 ~ u i ( x ) : (3) Let ~ u ( x; t ) = ( u ( x; t ) ; u ( x; t )) ; ~ f ( x; t ) = ( f ( x; t ) ; f ( x; t )) : and ~ u 0 ( x; t ) = ( u 0 ( x; t ) ; u 0 ( x; t )) ; ~ f 0 ( x; t ) = ( f 0 ( x; t ) ; f 0 ( x; t )) : Therefore, the related fuzzy inte gro-dif ferential equation (1) can be written as follo ws u 0 ( x; t ) = f ( x; t ) + Z x a k 1 ( x; s ) F 1 ( u ( s; t )) ds + Z a k 2 ( x; s ) F 2 ( u ( s; t )) ds (4) u 0 ( x; t ) = f ( x; t ) + Z x a k 1 ( x; s ) F 1 ( u ( s; t )) ds + Z a k 2 ( x; s ) F 2 ( u ( s; t )) ds (5) Similar to Remark 2.1, let u c ( x; t ) = u ( x; t ) + u ( x; t ) 2 ; u d ( x; t ) = u ( x; t ) u ( x; t ) 2 : (6) and f c ( x; t ) = f ( x; t ) + f ( x; t ) 2 ; f d ( x; t ) = f ( x; t ) f ( x; t ) 2 : (7) then (4) and (5) can be written as u 0 c ( x; t ) = f c ( x; t ) + Z x a k 1 ( x; s ) F 1 ( u c ( s; t )) ds + Z a k 2 ( x; s ) F 2 ( u c ( s; t )) ds (8) u 0 d ( x; t ) = f d ( x; t ) + Z x a k 1 ( x; s ) F 1 ( u d ( s; t )) ds + Z a k 2 ( x; s ) F 2 ( u d ( s; t )) ds (9) and u c (0 ; t ) = u (0 ; t ) + u (0 ; t ) 2 ; u d (0 ; t ) = u (0 ; t ) u (0 ; t ) 2 : (10) 4. HOMO T OPY AN AL YSIS METHOD (HAM) In this section, we shall describe the solution approaches bas ed on HAM for fuzzy V olterra-Fredholm inte gro- dif ferential equations. F or this, we consider the first equation of (8) namely we apply HAM for finding u c ( x; t ) ; and second equation approach is similar to the first one [22]. Consider u 0 c ( x; t ) = f c ( x; t ) + Z x a k 1 ( x; s ) F 1 ( u c ( s; t )) ds + Z a k 2 ( x; s ) F 2 ( u c ( s; t )) ds; (11) Indonesian J Elec Eng & Comp Sci V ol. 11, No. 3, September 2018: 857 -867 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 861 with initial condition u c (0 ; t ) : W e first construct the zero-order deformation equation (1 q )[ ' ( x; t ; q ) u c 0 ( x; t )] = q ~ N [ ' ( x; t ; q )] ; (12) subject to the initial condition ' (0 ; t ; q ) = u c 0 ( x; t ) = u c (0 ; t ) ; (13) where q 2 [0 ; 1] is the embedding parameter and ~ 6 = 0 is an auxiliary parameter and [ ' ( x; t ; q )] = @ [ ' ( x; t ; q )] @ x ; (14) with the property [ C ] = 0 ; (15) where C is inte gral constant. Also from (11), we can define N [ ' ( x; t ; q )] = @ [ ' ( x; t ; q )] @ x f c ( x; t ) Z x a k 1 ( x; s ) F 1 ( ' ( s; t ; q )) ds Z a k 2 ( x; s ) F 2 ( ' ( s; t ; q )) ds; (16) When parameter of q increases from 0 to 1 , then homotop y solution ' ( x; t ; q ) v aries from u c 0 ( x; t ) to solution u c ( t; r ) of the original equation (11). Using the parameter q ; ' ( x; t ; q ) can be e xpanded in T aylor series as follo ws ' ( x; t ; q ) = u c 0 ( x; t ) + 1 X m =0 u c m ( x; t ) q m ; where u c m ( x; t ) = 1 m ! @ m [ ' ( x; t ; q )] @ m q j q =0 : Assuming that auxiliary parameter ~ is properly selected so that the abo v e series is con v er gent when q = 1 ; then the solution u c ( x; t ) can be gi v en by u c ( x; t ) = u c 0 ( x; t ) + 1 X m =0 u c m ( x; t ) : Dif ferentiating (12) and initial condition (13) m-times with respect to q , then setting q = 0 ; and finally di viding them by m ! , we g ain the m th -order deformation equation [ u c m ( x; t ) m u c m 1 ( x; t )] = ~ < m ! ( u c m 1 ) ; (17) subject to the follo wing initial conditions, u c m (0 ; t ) = 0 ; (18) where < m ! ( u c m 1 ) = 1 ( m 1)! @ m 1 N [ ' ( x; t ; q )] @ m 1 q j q =0 (19) = @ u c m 1 ( x; t ) @ x (1 m ) f c ( x; t ) Z x a k 1 ( x; s ) F 1 ( u c m 1 ( s; t )) ds Z a k 2 ( x; s ) F 2 ( u c m 1 ( s; t )) ds; and m = ( 0 m 1 ; 1 m > 1 : Note that the high-order deformation Eq.(17) is go v erning the linear operator L , and the term < m ( ! u c m 1 ) can be e xpressed simply by Eq.(19) for an y nonlinear operator N : Similarly , we can apply HAM for finding u d ( x; t ) : Homotopy Analysis Method for the F ir st Or der Fuzzy V olterr a-F r edholm... (Ahmed A. Hamoud) Evaluation Warning : The document was created with Spire.PDF for Python.
862 ISSN: 2502-4752 5. MAIN RESUL TS A function ~ u is a solution of t he initial v alue problem (1)–(2) if and only if it is continuous and satisfies the inte gral equation ~ u ( x ) = ~ u 0 + Z x a ~ f ( t ) dt + Z x a Z t a k 1 ( t; s ) F 1 ( ~ u ( s )) dsdt + Z x a Z a k 2 ( t; s ) F 2 ( ~ u ( s )) dsdt; By changing the order of the inte gration, we ha v e ~ u ( x ) = ~ u 0 + Z x a ~ f ( t ) dt + Z x a Z x s k 1 ( t; s ) F 1 ( ~ u ( s )) dtds + Z x a Z a k 2 ( t; s ) F 2 ( ~ u ( s )) dsdt; ~ u ( x ) = ~ u 0 + Z x a ~ f ( t ) dt + Z x a Z x s k 1 ( t; s ) F 1 ( ~ u ( s )) dtds + Z a Z x a k 2 ( t; s ) F 2 ( ~ u ( s )) dtds; Since the function k 1 and k 1 are with no sign changes by assumption, we ha v e Z x s k 1 ( t; s ) F 1 ( ~ u ( s )) dt = Z x s k 1 ( t; s ) dt F 1 ( ~ u ( s )) Z x a k 2 ( t; s ) F 2 ( ~ u ( s )) dt = Z x a k 2 ( t; s ) dt F 2 ( ~ u ( s )) thus ~ u ( x ) = ~ g ( x ) + Z x a K 1 ( x; s ) F 1 ( ~ u ( s )) ds + Z a K 2 ( x; s ) F 2 ( ~ u ( s )) ds; (20) where ~ g ( x ) = ~ u 0 + Z x a ~ f ( t ) dt; K 1 ( x; s ) = Z x s k 1 ( t; s ) dt; K 2 ( x; s ) = Z x a k 2 ( t; s ) dt: Lemma 5..1 Let k 1 ( t; s ) be continuous in ( t; s ) and Lipsc hitz with r espect to s: Then K 1 ( x; s ) is Lipsc hitz with r espect to s: Proof. Let 0 s 1 s 2 x: Then j K 1 ( x; s 1 ) K 1 ( x; s 2 ) j = j Z x s 1 k 1 ( t; s 1 ) dt Z x s 2 k 1 ( t; s 2 ) dt j = j Z s 2 s 1 k 1 ( t; s 1 ) dt + Z x s 2 k 1 ( t; s 1 ) dt Z x s 2 k 1 ( t; s 2 ) dt j Z s 2 s 1 j k 1 ( t; s 1 ) j dt + Z x s 2 j k 1 ( t; s 1 ) k 1 ( t; s 2 ) j dt M j s 1 s 2 j + L j s 1 s 2 j ( x s 2 ) ( M + L ( a )) j s 1 s 2 j ; where M = max ( x;s ) 2 G j k 1 ( x; s ) j , G := f ( x; t ) j x 2 J ; t 2 [ a; x ] g J J ; and L is the Lipschitz constant of k 1 and thus K 1 satisfies in Lipschitz condition. Similarly , we can proof the procedure of K 2 ( x; s ) is Lipschitz with respect to s: Before starting and pro ving the main results, we introduce the follo wing h ypotheses: (A1) There e xist tw o constants M 1 ; M 2 > 0 such that, for an y u 1 ; u 2 2 C ( J ; R ) D " ( u c 1 ; u c 2 ) := sup x 2 J e "M 1 x D ( u c 1 ( s ) ; u c 2 ( s )) ; " 1 ; as a metric on X . (A2) There e xist tw o functions K 1 ; K 2 2 C ( D ; R + ) ; the set of all positi v e function continuous on D = f ( x; t ) 2 R R : 0 t x 1 g such that M 1 = max x;s 2 G j K 1 ( x; t ) j < 1 , M 2 = max x;s 2 G j K 2 ( x; t ) j < 1 ; Indonesian J Elec Eng & Comp Sci V ol. 11, No. 3, September 2018: 857 -867 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 863 (A3) The function g : J ! R is continuous. Theor em 5..1 Assume that (A1), (A2) and (A3) hold. If e "M 1 x e "M 1 a " + e "M 2 e "M 2 a " < 1 : Then the initial value pr oblem (1) - (2) has a unique solution. Proof. Let the op e rator A : X ! X : T o do this, it is e vident that ( Au c )( x ) 2 R F for all x 2 J and thus Au c : J ! R F be defined by ( Au c )( x ) = g c ( x ) + Z x a K 1 ( x; s ) F 1 ( u c ( s )) ds + Z a K 2 ( x; s ) F 2 ( u c ( s )) ds; D ( Au c 1 ( x ) ; Au c 2 ( x )) = D ( g c ( x ) + Z x a K 1 ( x; s ) F 1 ( u c 1 ( s )) ds + Z a K 2 ( x; s ) F 2 ( u c 1 ( s )) ds; g c ( x ) + Z x a K 1 ( x; s ) F 1 ( u c 2 ( s )) ds + Z a K 2 ( x; s ) F 2 ( u c 2 ( s )) ds ) ; = D ( Z x a K 1 ( x; s ) F 1 ( u c 1 ( s )) ds + Z a K 2 ( x; s ) F 2 ( u c 1 ( s )) ds; Z x a K 1 ( x; s ) F 1 ( u c 2 ( s )) ds + Z a K 2 ( x; s ) F 2 ( u c 2 ( s )) ds ) Z x a D ( K 1 ( x; s ) F 1 ( u c 1 ( s )) ; K 1 ( x; s ) F 1 ( u c 2 ( s ))) ds + Z a D ( K 2 ( x; s ) F 2 ( u c 1 ( s )) ; K 2 ( x; s ) F 2 ( u c 2 ( s ))) ds Z x a M 1 D ( u c 1 ( s ) ; u c 2 ( s )) ds + Z a M 2 D ( u c 1 ( s ) ; u c 2 ( s )) ds = Z x a M 1 e "M 1 s e "M 1 s D ( u c 1 ( s ) ; u c 2 ( s )) ds + Z a M 2 e "M 2 s e "M 2 s D ( u c 1 ( s ) ; u c 2 ( s )) ds Z x a M 1 e "M 1 s D " ( u c 1 ; u c 2 ) ds + Z a M 2 e "M 2 s D " ( u c 1 ; u c 2 ) ds e "M 1 x e "M 1 a " D " ( u c 1 ; u c 2 ) + e M 2 e "M 2 a " D " ( u c 1 ; u c 2 ) = e "M 1 x e "M 1 a " + e "M 2 e "M 2 a " D " ( u c 1 ; u c 2 ) : Since e "M 1 x e "M 1 a " + e "M 2 e "M 2 a " < 1 ; the operator A is a cont raction mapping. By the Banach fix ed point theorem we conclude that the initial v alue problem (1)-(2) has a unique solution. Similarly , we can proof the procedure of u d ( x; t ) . No w , we will discuss the con v er gence of HAM for Eq.(1) in the fuzzy case which di vides into tw o crisp inte gro-dif ferential equations as Eqs.(8)-(9). Theor em 5..2 Let the series P 1 m =0 u c m ( x; t ) con ver g e to u c ( x; t ) , wher e u c m ( x; t ) is pr oduced by the m -or der def o r - mation (17) , and besides P 1 m =0 u c m ( x; t ) con ver g es, then u c ( x; t ) is the e xact solution of V olterr a-F r edholm inte gr o- dif fer ential equation (8) when using HAM. Proof. W e assume P 1 m =0 u c m ( x; t ) con v er ge uniformly to u c ( x; t ) then lim m !1 u c m ( x; t ) = 0 ; 8 0 x ; 0 t 1 : Homotopy Analysis Method for the F ir st Or der Fuzzy V olterr a-F r edholm... (Ahmed A. Hamoud) Evaluation Warning : The document was created with Spire.PDF for Python.
864 ISSN: 2502-4752 W e can write, n X m =1 [ u c m ( x; t ) m u c m 1 ( x; t )] = u c 1 ( x; t ) + ( u c 2 ( x; t ) u c 1 ( x; t )) +( u c 3 ( x; t ) u c 2 ( x; t )) + : : : +( u c n ( x; t ) u c n 1 ( x; t )) = u c n ( x; t ) : (21) Hence, from Eq.(21) lim n !1 u c n ( x ) = 0 : (22) So, using Eq.(22), we ha v e 1 X m =1 [ u c m ( x; t ) m u c m 1 ( x; t )] = 1 X m =1 [ u c m ( x; t ) m u c m 1 ( x; t )] = 0 : Therefore from Eq.(22), we can obtain that 1 X m =1 [ u c m ( x; t ) m u c m 1 ( x; t )] = ~ 1 X m =1 < m 1 ( ! u c m 1 ( x; t )) = 0 : Since ~ 6 = 0 and we ha v e 1 X m =1 < m 1 ( ! u c m 1 ( x; t )) = 0 : (23) By substituting < m 1 ( ! u c m 1 ) into the relation (23) and simplifying it, we ha v e < m 1 ( ! u c m 1 ( x; t )) = 1 X m =1 [ @ u c m 1 ( x; t ) @ x Z x a k 1 ( x; s ) F 1 ( u c m 1 ( s; t )) ds Z a k 2 ( x; s ) F 2 ( u c m 1 ( s; t )) ds (1 m ) f c ( x; t )] ; = ( 1 X m =1 @ u c m 1 ( x; t ) @ x Z x a k 1 ( x; s )[ 1 X m =1 F 1 ( u c m 1 ( s; t ))] ds Z a k 2 ( x; s )[ 1 X m =1 F 2 ( u c m 1 ( s; t ))] ds 1 X m =1 (1 m ) f c ( x; t ) ; = @ u c ( x; t ) @ x Z x a k 1 ( x; s ) F 1 ( u c ( s; t )) ds Z a k 2 ( x; t ) F 2 ( u c ( s; t )) ds f c ( x; t ) : (24) From Eq.(23) and Eq.(24), we ha v e @ u c ( x; t ) @ x = f c ( x; t ) + Z x a k 1 ( x; s ) F 1 ( u c ( s; t )) ds + Z a k 2 ( x; s ) F 2 ( u c ( s; t )) ds; therefore, u c ( x; t ) must be the e xact solution of Eq.(8). Similarly , we can proof the procedure of u d ( x; t ) . Then, u ( x; t ) must be the e xact solution of Eq.(1), and the proof is complete. 6. ILLUSTRA TIVE EXAMPLE In this section, we present the analytical technique based on HAM to solv e fuzzy V olterra-Fredholm inte gro- dif ferential equation. Indonesian J Elec Eng & Comp Sci V ol. 11, No. 3, September 2018: 857 -867 Evaluation Warning : The document was created with Spire.PDF for Python.
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752 865 Example 1. Let us consider fuzzy V olterra-Fredholm inte gro-dif ferential equation: @ ~ u ( x; t ) @ x + ~ u ( x; t ) = ~ f ( x; t ) + Z x 0 xs 2 ~ u ( s; t ) ds + Z 1 0 xs ~ u ( s; t ) ds; (25) where ~ u ( x; t ) = ( u ( x; t ) ; u ( x; t )) ; ~ f ( x; t ) = ( f ( x; t ) ; f ( x; t )) f ( x; t ) = t 2 x 3 + 1 ; f ( x; t ) = 2 tx 3 4 x t + 8 ; = 0 ; = 1 : with initial conditions u (0 ; t ) = 0 ; u (0 ; t ) = 8 : Exact solution of this fuzzy V olterra-Fredholm inte gro-dif ferential equation is gi v en by ~ u ( x; t ) = ( xt; 8 xt ) : From Eq.(7), we ha v e f c ( x; t ) = f ( x; t ) + f ( x; t ) 2 = 4 2 x; f d ( x; t ) = f ( x; t ) f ( x; t ) 2 = 4 t 2 x 2 tx 3 : The e xact solution of related crisp equations are as follo ws u c ( x; t ) = u ( x; t ) + u ( x; t ) 2 = 4 ; u d ( x; t ) = u ( x; t ) u ( x; t ) 2 = 4 tx: From Eqs.(16) and(25) can be written N [ ' ( x; t ; q )] = @ [ ' ( x; t ; q )] @ x + ' ( x; t ; q ) 4 + 2 x Z 1 0 k 2 ( x; s ) F 2 ( ' ( s; t ; q )) ds; No w , using m th -order deformation equation and initial conditions, we recursi v ely obtain u c 0 ( x; t ) = u c (0 ; t ) = u (0 ; t ) + u (0 ; t ) 2 = 4 u c 1 ( x; t ) = 0 ; u c 2 ( x; t ) = 0 ; : : : Thus the approximate HAM solution Y c m ( x; t ) = m X n =0 u c n ( x; t ) = 4 The approximate solution same as e xact solution. Similarly , to approximate u d ( t; r ) ; N [ ' ( x; t ; q )] = @ [ ' ( x; t ; q )] @ x + ' ( x; t ; q ) 4 + t + 2 x + 2 tx 3 Z 1 0 k 2 ( x; s ) F 2 ( ' ( s; t ; q )) ds; No w , using m th -order deformation equation and initial conditions, we recursi v ely obtain u d 0 ( x; t ) = u d (0 ; t ) = u (0 ; t ) u (0 ; t ) 2 = 4 ; u d 1 ( x; t ) = ~ 3 xt (3 + x ) ; u d 2 ( x; t ) = ~ 72 xt ( ~ (72 + 45 x + 8 x 2 ) + 72 + 24 x ) ; : : : Homotopy Analysis Method for the F ir st Or der Fuzzy V olterr a-F r edholm... (Ahmed A. Hamoud) Evaluation Warning : The document was created with Spire.PDF for Python.
866 ISSN: 2502-4752 Thus the approximate HAM solution when ~ = 1 Y d m ( x; t ) = m X n =0 u d n ( x; t ) = u d 0 ( x; t ) + u d 1 ( x; t ) + + u d m ( x; t ) 4 tx: Note that, we can control the con v er gence re gion of HAM series solution by the auxiliary parameter ~ . 7. CONCLUSION Homotop y analysis method has been performed to find approximate analytical s olutions for fuzzy V olterra- Fredholm inte gro-dif ferential equations. The reliability of the method and reduction in the size of the computational w ork gi v e this method a wider applicability . The method is v ery po werful and ef ficient in finding analytical as well as numerical solutions for wide classes of linear and nonlinear fuzzy inte gro-dif ferent ial equations. Obtained results sho w that we can control the con v er gence re gion of HAM series solution by the auxiliary parameter ~ . The illustrati v e e xample and con v er gence theorem sho w the ef ficienc y and accurac y of the HAM. REFERENCES [1] S. Chang and L. Zadeh, ”On fuzzy mapping and control, IEEE T r ans. Systems, Man cybernet, (1972), 2, pp. 30–34. [2] D. Dubois and H. Prade, ”Operations on fuzzy numbers, Int. J . systems Science , (1978), 9, pp. 613–626. [3] D. Dubois and H. Prade, ”T o w ards fuzzy dif ferential calculus. Inte grati on of fuzzy mappings, Fuzzy Sets and Systems, (1982), 8, pp. 1–17. [4] O. Kale v a, ”Fuzzy dif ferential equation, Fuzzy Sets and Systems, (1987), 24, pp. 301–317. [5] S. Narayanamoorth y and S. Sathiyapriya, ”A pertinent approach to s olv e nonlinear fuzzy inte gro-dif ferential equa- tions, Spring er Plus. (2016), 5, pp. 1–17. [6] T . Allahviranloo, S. Abbasbandy and S. Hashemzehi, ”Approximating the solution of the linear and nonlinear fuzzy V olterra inte gro-dif ferential equations using e xpansion method, Abstr . Appl. Anal. (2014), pp. 1–8. [7] M. Das and D. T alukdar , ”Method for solving fuzzy inte gro-dif ferential equations by us ing fuzzy Laplace trans- formation, Int. J . Sci. T ec h. (2014), 5, pp. 291–295. [8] E. Hussain and A. Ali, ”Homotop y analysis method for solving fuzzy int e gro-di f ferential equations, Modern Appl. Sci. (2013), 7, pp. 15–25. [9] S. Behiry and S. Mohamed, ”Solving high-order nonlinear V olterra–Fredholm inte gro-dif ferential equations by dif ferential transform method, Nat. Sci. (2012), 8, pp. 581–587. [10] H. Rahimi, M. Khezerloo and S. Khezerloo, ”Approximating the fuzzy solution of the non-linear fuzzy V olterra inte gro-dif ferential equation using fix ed point theorems, Int J Indus Math. (2011), 3(3), pp. 227–236. [11] M. Hashemi and S. Abbas bandy , ”The solution of fuzzy V olterra-Fredholm inte gro-dif ferential equations using v ariational iteration method, Int. J . Math. Comput. (2011), 11, pp. 29–38. [12] A. Hamoud and K. Ghadle, ”modified Adomian decomposition method for s olving fuzzy V olterra-Fredholm inte gral equations, J ournal of the Indian Math. Soc. (2018), 85(1-2), pp. 01–17. [13] A. Hamoud and K. Ghadle, ”Existence and uniqueness of solutions for fractional mix ed V olterra-Fredholm inte gro-dif ferential equations, Indian J. Math. (2018), 60(3), pp. 375–395. [14] A. Hamoud, K. Ghadle, M. Bani Issa and Ginisw amy , ”Existence and uniqueness t heorems for fractional V olterra-Fredholm inte gro-dif ferential equations, Int. J. Appl. Math. (2018), 31(3), pp. 333–348. [15] A. Hamoud, K. Ghadle and S. Atshan, ”The approximate solutions of frac tional inte gro-dif ferential equations by using modified Adomian decomposition method, Khayyam J. Math. (2019), 5(1), pp. 21–39. [16] A. Hamoud and K. Ghadle, ”Existence and uniqueness of the solution for V olterra-Fredholm inte gro-dif ferential equations, Journal of Siberian Federal Uni v ersity . Mathematics & Ph ysics , (2018), 11(6), pp. 692–701. [17] A. Hamoud and K. Ghadle, ”Some ne w e xistence, uniqueness and con v er gence results for fractional V olterra- Fredholm inte gro-dif ferential equations, J. Appl. Comput. Mech. (2019), 5(1), pp. 58–69. [18] A. Hamoud, M. Bani Issa and K. Ghadle, ”Existence and uniqueness results for nonlinear V olterra-Fredholm inte gro-dif ferential equations, Nonlinear Functional Analysis and Applications , (2018), 23(4), pp. 797–805. [19] A. Hamoud and K. Ghadle, ”Usage of the homotop y analysis method for solving fractional V olterra-Fredholm inte gro-dif ferential equation of the second kind, T amkang Journal of Mathematics , (2018), 49(4), pp. 301–315. [20] A. Hamoud and K. Ghadle, ”Recent adv ances on reliable methods for solving V olterra-Fredholm inte gral and inte gro-dif ferential equations, Asian J. Math. Comput. Res. (2018), 24, pp. 128–157. [21] Y . Chalco-Cano and H. Roman-Flores, ”On ne w solutions of fuzzy dif ferential equations, Chaos, Solitons and F r actals , (2008), 38, pp. 112–119. Indonesian J Elec Eng & Comp Sci V ol. 11, No. 3, September 2018: 857 -867 Evaluation Warning : The document was created with Spire.PDF for Python.