EUROPEAN JOURNAL OF PURE AND APPLIED MATHEMATICS Vol. 13, No. 3, 2020, 444-458 ISSN 1307-5543 – www.ejpam.com Published by New York Business Global On Multi Poly-Genocchi Polynomials with Parameters a, b and c Roberto B. Corcino1,∗, Mark P. Laurente2, Mary Ann Ritzell P. Vega3 1 Research Institute for Computational Mathematics and Physics, Cebu Normal University, 6000 Cebu City, Philippines 2 Department of Mathematics, Mindanao State University, Marawi City, Philippines 3 Department of Mathematics and Statistics, College of Science and Mathematics, Mindanao State University-Iligan Institute of Technology, 9200 Iligan City, Philippines Abstract. Most identities of Genocchi numbers and polynomials are related to the well-known Benoulli and Euler polynomials. In this paper, multi poly-Genocchi polynomials with parameters a, b and c are defined by means of polylogarithm in multiple paramaters. Several properties of these polynomials are established including some recurrence relations and explicit formulas. 2020 Mathematics Subject Classifications: 11B68, 11B73, 05A15 Key Words and Phrases: Bernoulli numbers, Euler numbers, Genocchi numbers, poly- Bernoulli numbers, poly-Euler numbers poly-Genocchi numbers 1. Introduction The Genocchi numbers and polynomials can be traced back to Angelo Genocchi (1817- 1889). Genocchi numbers have been extensively studied in many different contexts in mathematics. For instance, Genocchi numbers have been studied by several authors in the context of Apostol-type polynomials, Hermite-type polynomials, polylogarithm, and their q-analogues [3, 6–8, 13, 20, 21, 23, 27, 29, 30]. Many studies and literature provide relations of Genocchi numbers to Bernoulli and Euler numbers, especially Euler numbers. Bernoulli, Euler and Genocchi numbers defined by exponential generating function (see [1, 19, 22]) ∞∑ n=0 Bn tn n! = t et − 1 , |t| < 2π (1) ∞∑ n=0 En tn n! = 2 et + 1 , |t| < π (2) ∗Corresponding author. DOI: https://doi.org/10.29020/nybg.ejpam.v13i3.3676 Email addresses: rcorcino@yahoo.com (R. Corcino), mark.laurente2012@gmail.com (M. Laurente), maryannritzel.vega@g.msuiit.edu.ph (MAR. Vega) https://www.ejpam.com 444 c© 2020 EJPAM All rights reserved. R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 445 ∞∑ n=0 Gn tn n! = 2t et + 1 , |t| < π. (3) The Bernoulli, Euler and Genocchi polynomials are defined via generating functions to be, respectively, ∞∑ n=0 Bn(x) tn n! = t et − 1 ext, |t| < 2π (4) ∞∑ n=0 En(x) tn n! = 2 et + 1 ext, |t| < π (5) ∞∑ n=0 Gn(x) tn n! = 2t et + 1 ext, |t| < π, (6) where, when x = 0, Bn(0) = Bn, En(0) = En and Gn(0) = Gn. (see [5, 19, 22, 24]) Araci [19] and Kim et al. [26] did some researches on the so-called Genocchi polyno- mials of higher order arising from Genocchi basis, which were defined by( 2t et + 1 )k ext = ∞∑ n=0 G(k) n (x) tn n! . (7) The main objective of their studies is to derive interesting identities on (7) using a new method constructed by Kim et al. [11]. Moreover, Araci and He [7, 19, 30] introduced the Apostol-Genocchi polynomials as an extension of the Genocchi polynomials, which were defined by 2t λet + 1 ext = ∞∑ n=0 Gn(x, λ) tn n! . Based on this, Araci [23] introduced Apostol-Genocchi polynomials of higher order which is also called the generalized Apostol-Genocchi polynomials of order k ∈ C,( 2t λet + 1 )k ext = ∞∑ n=0 G(k) n (x, λ) tn n! , |t| < π when λ = 1 and (8) |t| < | log(−λ)| when λ 6= 1;λ ∈ C. In [15], Lim defined the degenerated Genocchi polynomials G (k) n (x, λ) of order k to be( 2t (1 + λt)1/λ )k (1 + λt)x/λ = ∞∑ n=0 G(k) n (x, λ) tn n! . Besides these generalizations, Araci [20], Duran et al. [29] and Agyuz et al. [6] also introduced the q-analogue of the Genocchi polynomials as follows, ∞∑ n=0 Gn,q(x) tn n! = t ∫ zp q−yet[y+x]qdµ−q(y), R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 446 where [x]q = 1− qx 1− q , [x]−q = 1− (−q)x 1 + q . This definition is constructed by p-adic fermionic q-integral on Zp with respect to µ−q. It can also be defined by ∞∑ n=0 Gn,q(x) tn n! = [2]qt ∞∑ m=0 (−1)met[m+x]q . In which when we take x = 0, it becomes Gn,q(0) := Gn,q, which we call it the nth q-Genocchi number. When it comes to Genocchi numbers, the most common thing that comes to our mind is to determine the relationship between Genocchi numbers, Bernoulli numbers and Euler numbers [22]. Indeed, most researches on Genocchi numbers concern the relations between these three kinds of numbers [3, 4, 12, 22]. In other words, there are many literatures that provide identities on these three kinds of numbers. Similarly, when it comes to Genocchi polynomials, the most common thing is to establish relationship between Genocchi polynomials, Bernoulli polynomials and Euler polynomials [2–4, 8, 12, 22, 30]. Another form of generalization of Bernoulli polynomials was introduced by Kaneko [10]. This generalization was defined in terms of the following kth polylogarithm Lik(z): Lik(z) = ∞∑ n=1 zn nk . (9) where k ∈ Z and z ∈ C with |z| < 1 which can be extended to z ≥ 1 by the process of analytic continuation. When z = 1, the kth polylogarithm gives the Riemann zeta function. That is, Lik(1) = ζ(k) = ∞∑ n=1 1 nk . Also, when k = 1, the 1st polylogarithm yields the natural logarithmic function as follows: Li1(z) = − ln(1− z). This special case of the polylogarithm motivates the construction of poly-Bernoulli num- bers in the sense that Li1(1− e−x) = x. The poly-Bernoulli numbers B (k) n were defined by Kaneko [10] as Lik(1− e−t) et − 1 = ∞∑ n=0 B(k) n tn n! . (10) R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 447 Parallel to this, Kim et al. [27] defined poly-Genocchi polynomials as follows 2Lik(1− e−t) et + 1 ext = ∞∑ n=0 G(k) n (x) tn n! . (11) Note that, when x = 0, (11) reduces to 2Lik(1− e−t) et + 1 = ∞∑ n=0 G(k) n tn n! . (12) where G (k) n are called the poly-Genocchi numbers. Moreover, they defined a modified poly-Genocchi polynomials, denoted by G (k) n,2(x), as follows Lik(1− e−2t) et + 1 ext = ∞∑ n=0 G (k) n,2(x) tn n! . (13) Note that G(1) n := G (1) n,2 := Gn(x). Kim et al. [27] obtained several properties of these polynomials. On the other hand, Kurt [13] defined two forms of generalized poly-Genocchi polyno- mials with parameters a, b, and c, as follows 2Lik(1− (ab)−t) a−t + bt ext = ∞∑ n=0 G(k) n (x; a, b, c) tn n! (14) 2Lik(1− (ab)−2t) a−t + bt ext = ∞∑ n=0 G (k) n,2(x; a, b, c) tn n! , (15) which are motivated by the definitions in (11) and (13), respectively. Kurt [13] also derived several properties parallel to those of poly-Genocchi polynomials by Kim et al. [27]. The followings are some relations between poly-Bernoulli and poly-Genocchi numbers and polynomials; poly-Genocchi numbers, Euler number and Stirling numbers of the sec- ond kind; and modified poly-Bernoulli and poly-Genocchi polynomials: nB (k) n−1 = 1 2 nG (k) n−1 + n∑ m=0 ( n m ) BmG (k) n−m (16) B (k) n,2 − 2n+1B(k) n = 1 2 G (k) n,2 (17) 2nG (k) n−1 − 2 n∑ m=0 ( n m ) GmG (k) n−m = n∑ m=0 ( n m ) Gm(1)G (k) n−m, (18) R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 448 n∑ m=0 ( n m ) B(k) m (x)B (k) n−m(y) = n∑ p=0 ( n p ) B(k) p B (k) n−p(x+ y), (19) and G (k) n,2 = 2n+1 ( B(k) n ( x+ 1 2 ) −B(k) n (x) ) . (20) Moreover, using the generating function of the poly-Genocchi numbers and Stirling numbers of the second kind, we have ∞∑ n=0 G(k) n tn n! = 2 et + 1 ∞∑ m=1 (−1)m(e−t − 1)m mk = ∞∑ m=1 (−1)m mk 2 et + 1 m! ∞∑ l=0 S2(l,m)(−1)l tl l! = ∞∑ m=1 (−1)m mk ∞∑ n=0 En tn n! m! ∞∑ l=0 S2(l,m)(−1)l tl l! = ∞∑ m=1 ∞∑ n=0 ∞∑ l=0 (−1)m mk Enm!S2(l,m)(−1)l tn+l n!l! . Replacing n+ l with l, we get ∞∑ n=0 G(k) n tn n! = ∞∑ m=1 ∞∑ n=0 ∞∑ l=n (−1)m mk Enm!S2(l − n,m)(−1)l−n tl n!(l − n)! l! l! = ∞∑ m=1 ∞∑ n=0 ∞∑ l=n ( l n ) (−1)m+l−n mk Enm!S2(l − n,m) tl l! = ∞∑ l=0 { l∑ n=0 ∞∑ m=1 ( l n ) (−1)m+l−n mk Enm!S2(l − n,m) } tl l! = ∞∑ n=0 { n∑ r=0 ∞∑ m=1 ( n r ) (−1)m+n−r mk Erm!S2(n− r,m) } tn n! . By comparing the coefficient of tn n! , we obtain G(k) n = n∑ r=0 ( n r ){ ∞∑ m=1 (−1)m+n−r mk Erm!S2(n− r,m) } . (21) The multi poly-Bernoulli numbers was first introduced by Imatomi et al. [9] using the concept of multiple polylogarithm also known as multiple zeta values, which is given by Li(k1,k2,...,kr)(z) = ∑ 00 zm1 mk1 1 , which is exactly (9). The multi poly-Bernoulli numbers defined by Imatomi et al. [9] is given by Li(k1,k2,...,kr)(1− e−t) 1− e−t = ∞∑ n=0 B(k1,k2,...,kr) n tn n! . These numbers possess the following recurrence relation and explicit formula B(k1,k2,...,kr) n = 1 n+ 1 ( B(k1−1,k2,...,kr) n − n−1∑ m=1 ( n m− 1 ) B(k1,k2,...,kr) m ) B(k1,k2,...,kr) n = (−1)n ∑ n+1≥m1>m2>...>mr>0 (−1)m1−1(m1 − 1)!S(n,m1 − 1) mk1 1 m k2 2 ...m kr r . Parallel to the above generalization is the generalized multi poly-Bernoulli polynomials which are denoted by B (k1,k2,...,kr) n (x; a, b, c). These polynomials have been introduced in [18] by means of the above multiple poly-logarithm. More precisely, we have Li(k1,k2,...,kr)(1− (ab)−t) (a−t − bt)r crxt = ∞∑ n=0 B(k1,k2,...,kr) n (x; a, b, c) tn n! . (23) When r = 1, (23) boils down to the generalized poly-Bernoulli polynomials with three parameters a, b, c. Moreover, when c = e, (23) reduces to the multi poly-Bernoulli poly- nomials with two parameters a, b. These special cases have been discussed intensively in [18]. On the other hand, the generalized multi poly-Euler polynomials were also defined in [17] by means of multiple poly-logarithm. This paper intends to investigate multi poly-Genocchi polynomials with parameters a, b and c. 2. Multi Poly-Genocchi Polynomials with Parameters a, b and c In this section, using the concept of multiple polylogarithm, we introduce the multi poly-Genocchi polynomials with parameters a, b and c. Some properties of these polyno- mials are established parallel to those of the poly-Genocchi polynomials with parameters a, b and c. Definition 2.1. The multi poly-Genocchi polynomials with parameters a, b and c are defined by ∞∑ n=0 G(k1,k2,...,kr)n (x; a, b, c) tn n! = Li(k1,k2,...,kr)(1− (ab)−2t) (a−t + bt)r crxt (24) R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 450 When c = e, equation (24) reduces to ∞∑ n=0 G(k1,k2,...,kr)n (x; a, b, e) tn n! = Li(k1,k2,...,kr)(1− (ab)−2t) (a−t + bt)r erxt. For convenience, we use G(k1,k2,...,kr)n (x; a, b) to denote G(k1,k2,...,kr)n (x; a, b, e). That is, ∞∑ n=0 G(k1,k2,...,kr)n (x; a, b) tn n! = Li(k1,k2,...,kr)(1− (ab)−2t) (a−t + bt)r erxt. (25) Furthermore, if we put a = 1, b = e in (25), then this will reduce to ∞∑ n=0 G(k1,k2,...,kr)n (x; 1, e) tn n! = Li(k1,k2,...,kr)(1− e−2t) (1 + et)r erxt. We use G(k1,k2,...,kr)n (x) to denote G(k1,k2,...,kr)n (x; 1, e). That is, ∞∑ n=0 G(k1,k2,...,kr)n (x) tn n! = Li(k1,k2,...,kr)(1− e−2t) (1 + et)r erxt. (26) When x = 0, equation (25) gives ∞∑ n=0 G(k1,k2,...,kr)n (0; a, b) tn n! = Li(k1,k2,...,kr)(1− (ab)−2t) (a−t + bt)r . We use G(k1,k2,...,kr)n (a, b) to denote G(k1,k2,...,kr)n (0; a, b). The following theorem is given without proof since it follows from [16, Theorems 2.1- 2.3]. Theorem 2.2. The generalized poly-Genocchi polynomials satisfy the relations G(k1,k2,...,kr)n (x; a, b, c) = (r(ln a+ ln b))nG(k1,k2,...,kr)n ( x ln c+ ln a ln ab ) (27) G(k1,k2,...,kr)n (x; a, b, c) = ∞∑ i=0 ( n i ) (r ln c)n−iG(k1,k2,...,kr)i (a, b)xn−i (28) d dx G(k1,k2,...,kr)n+1 (x; a, b, c) = (n+ 1)(r ln c)G(k1,k2,...,kr)n (x; a, b, c). (29) Equation (29) contains a differential identity that can be used to classify generalized poly-Genocchi polynomials as Appell polynomials [14, 25, 28]. When c = e1/r, equation (29) reduces to d dx G(k1,k2,...,kr)n+1 (x; a, b, e1/r) = (n+ 1)G(k1,k2,...,kr)n (x; a, b, e1/r), (30) R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 451 which is one of the property for the polynomial to be classified as Appell polynomial. Hence, the generalized poly-Genocchi polynomials G(k)n (x; a, b) must possess the following properties G(k1,k2,...,kr)n (x; a, b, e1/r) = n∑ i=0 ( n i ) cix n−i G(k1,k2,...,kr)n (x; a, b, e1/r) = ( ∞∑ i=0 ci i! Di ) xn, for some scalar ci 6= 0. It is then necessary to find the sequence {cn}. However, using equation (28), ci = G(k1,k2,...,kr)i (a, b), which implies the following corollary. Corollary 2.3. The generalized poly-Genocchi polynomials satisfy the following formula G(k1,k2,...,kr)n (x; a, b, e1/r) = ( ∞∑ i=0 G(k1,k2,...,kr)i (a, b) i! Di ) xn. For example, when n = 3, we have G(k1,k2,...,kr)3 (x; a, b, e1/r) = ( ∞∑ i=0 G(k1,k2,...,kr)i (a, b) i! Di ) x3 = G(k1,k2,...,kr)0 (a, b) 0! x3 + G(k1,k2,...,kr)1 (a, b) 1! D1x3 + G(k1,k2,...,kr)2 (a, b) 2! D2x3 + G(k1,k2,...,kr)3 (a, b) 3! D3x3 = G(k)0 (a, b)x3 + 3G(k1,k2,...,kr)1 (a, b)x2 + 3G(k1,k2,...,kr)2 (a, b)x+ G(k1,k2,...,kr)3 (a, b). The following theorem contains the addition formula for G(k1,k2,...,kr)n (x; a, b, c). Theorem 2.4. The generalized poly-Genocchi polynomials satisfy the following addition formula G(k1,k2,...,kr)n (x+ y; a, b, c) = ∞∑ i=0 ( n i ) (r ln c)n−iG(k1,k2,...,kr)i (x; a, b, c)yn−i . Proof. Using Definition 2.1, ∞∑ n=0 G(k1,k2,...,kr)n (x+ y; a, b, c) tn n! = Lik(1− (ab)−2t) (a−t + bt) r c(x+y)rt R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 452 = Li(k1,k2,...,kr)(1− (ab)−2t) (a−t + bt) r cxrtcyrt = ( ∞∑ n=0 G(k1,k2,...,kr)n (x; a, b, c) tn n! ) ( ∞∑ n=0 (yr ln c)n tn n! ) = ∞∑ n=0 ( ∞∑ i=0 ( n i ) (yr ln c)n−iG(k1,k2,...,kr)i (x; a, b, c) ) tn n! Comparing the coefficients of tn n! yields the desired result. When y = 1, Theorem 2.4 yields the following recurrence relation G(k1,k2,...,kr)n (x+ 1; a, b, c) = n∑ m=0 ( n m ) (r ln c)mG(k1,k2,...,kr)n−m (x; a, b, c). (31) The following corollary immediately follows from equation (36) and the characteriza- tion of Appell polynomials [14, 25, 28]. Corollary 2.5. The generalized poly-Genocchi polynomials satisfy the following addition formula G(k1,k2,...,kr)n (x+ y; a, b, e1/r) = ∞∑ i=0 ( n i ) G(k1,k2,...,kr)i (x; a, b, e1/r)yn−i. (32) Taking x = 0 in formula (32) and using the fact G(k)n (0; a, b, c) = G(k)n (a, b), Theorem 2.4 gives formula (28). The next theorem contains an expression of generalized poly-Genocchi polynomials in terms of multiple parameters poly-Bernoulli polynomials. Theorem 2.6. The of generalized poly-Genocchi polynomials satisfy the relation G(k1,k2,...,kr)n (x; a, b, c) = r∑ m=0 ( r m ) (−1)mB(k1,k2,...,kr) n ( rx ln c+ (r +m) ln a+m ln b 2(ln a+ ln b) ) (2 ln ab)n. (33) Proof. We can rewrite equation (24) as ∞∑ n=0 G(k1,k2,...,kr)n (x; a, b, c) tn n! = Li(k1,k2,...,kr)(1− (ab)−2t) (1− (ab)2t)r (e−t ln a−et ln b)rerxt ln ce2rt ln a = Li(k1,k2,...,kr)(1− e−2t(ln ab)) (−1)r(e2t ln(ab) − 1)r r∑ m=0 ( r m ) (−1)me(r−m)t(− ln a+x ln c+2 ln a)emt(ln b+x ln c+2 ln a) R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 453 = r∑ m=0 ( r m ) (−1)m Li(k1,k2,...,kr)(1− e−2t(ln ab)) (−1)r(e2t ln(ab) − 1)r e((rx ln c+(r+m) ln a+m ln b)/2 ln ab)(2t ln ab). Using the definition of poly-Bernoulli polynomials, we have ∞∑ n=0 G(k1,k2,...,kr)n (x; a, b, c) tn n! = ∞∑ n=0 { r∑ m=0 ( r m ) (−1)mB(k1,k2,...,kr) n ( rx ln c+ (r +m) ln a+m ln b 2(ln a+ ln b) ) 2n(ln a+ ln b)n } tn n! . Comparing the coefficients of tn n! yields (33). The next two theorems are given without proof since it follows from [16, Theorem 3.4 and Theorem 2.6]. Theorem 2.7. The generalized multi-poly-Genocchi polynomials have the following ex- plicit formula G(k1,k2,...,kr)n (x; a, b, c) = n∑ i=0 ∑ 0≤m1≤m2≤···≤mr c1+c2+···=r mr∑ j=0 (rx ln c− 2j ln ab)n−ir!(−1)j+s(s ln ab+ r ln a)i ( mr j )( n j ) (c1!c2! · · · )(mk1 1 m k2 2 · · ·m kr k ) (34) where s = c1 + 2c2 + · · · Theorem 2.8. The generalized poly-Genocchi polynomials G(k1,k2,...,kr)n (x; a, b) satisfy the following explicit formulas G(k1,k2,...,kr)n (x; a, b, c) = ∞∑ m=0 n∑ l=m { l m }( n l ) (ln c)lG(k1,k2,...,kr)n−l (−m ln c; a, b)(x)(m) (35) G(k1,k2,...,kr)n (x; a, b, c) = ∞∑ m=0 n∑ l=m { l m }( n l ) (ln c)lG(k1,k2,...,kr)n−l (a, b)(x)m (36) G(k1,k2,...,kr)n (x; a, b, c) = n∑ l=0 n−l∑ m=0 ( n l ){ l + s s }(n−l m )( l+s s )G(k1,k2,...,kr)n−l−m (a, b)B(s) m (x ln c) (37) G(k1,k2,...,kr)n (x; a, b, c) = ∞∑ m=0 ( n m ) (1− λ)s s∑ j=0 ( s j ) (−λ)s−jG(k1,k2,...,kr)n−m (j; a, b)H(s) m (x;λ), (38) where (x)(n) = x(x+ 1) · · · (x+ n− 1), (x)n = x(x− 1) · · · (x− n+ 1),( t et − 1 )s ext = ∞∑ n=0 B(s) n (x) tn n! and ( 1− λ et − λ )s ext = ∞∑ n=0 H(s) n (x;λ) tn n! . R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 454 3. Symmetrized Generalization In this section, we will consider the symmetrized generalization of multi poly-Genocchi polynomials with parameters a, b and c. Definition 3.1. For m,n ≥ 0, we define the symmetrized generalization of multi poly- Genocchi polynomials with parameters a, b and c as follows S(m) n (x, y; a, b, c) = ∑ k1+k2+...+kr=m ( m k1, k2, . . . kr ) G(−k1,−k2,...−kr−1) n (x; a, b, c) (ln a+ ln b)n ( (r − 1)y ln c+ ln a ln a+ ln b )kr . (39) The following theorem contains the double generating function for S(m) n (x, y; a, b, c). Theorem 3.2. For n,m ≥ 0, we have ∞∑ n=0 ∞∑ m=0 S(m) n (x, y; a, b, c) tn n! um m! = e ( (r−1)y ln c+ln a ln a+ln b ) u e (r−1) ( (r−1)x ln c+ln a ln a+ln b ) t e( r 2)u+2(r−1)t(1− e−2t)r−1 (1 + et)r−1 ∏r−1 i=1 (e2t + eiu − e2t+iu) . (40) Proof. ∞∑ n=0 ∞∑ m=0 S(m) n (x, y; a, b, c) tn n! um m! = ∞∑ n=0 ∞∑ m=0 ∑ k1+k2+...+kr=m G(−k1,−k2,...−kr−1) n (x; a, b, c) (ln a+ ln b)n ( (r − 1)y ln c+ ln a ln a+ ln b )kr tn n! × × um k1!k2! . . . kr! = ∞∑ n=0 ∑ k1+k2+...+kr≥0 G(−k1,−k2,...−kr−1) n (x; a, b, c) (ln a+ ln b)n ( (r − 1)y ln c+ ln a ln a+ ln b )kr tn n! × ×u k1+k2+...+kr k1!k2! . . . kr! = ∞∑ n=0 ∑ k1+k2+...+kr−1≥0 G(−k1,−k2,...−kr−1) n (x; a, b, c) (ln a+ ln b)n ∑ kr≥0 ( (r − 1)y ln c+ ln a ln a+ ln b )kr ukr kr! × × t n n! uk1+k2+...+kr−1 k1!k2! . . . kr−1! = e ( (r−1)y ln c+ln a ln a+ln b ) u ∞∑ n=0 ∑ k1+k2+...+kr−1≥0 G(−k1,−k2,...−kr−1) n (x; a, b, c) (ln a+ ln b)n tn n! uk1+k2+...+kr−1 k1!k2! . . . kr−1! R. Corcino, M. Laurente, MAR. Vega / Eur. J. Pure Appl. Math, 13 (3) (2020), 444-458 455 Using identity (26), we obtain ∞∑ n=0 ∞∑ m=0 S(m) n (x, y; a, b, c) tn n! um m! = e ( (r−1)y ln c+ln a ln a+ln b ) u ∑ k1+k2+...+kr−1≥0 ∞∑ n=0 G(−k1,−k2,...−kr−1) n ( (r − 1)x ln c+ ln a ln a+ ln b ) tn n! × ×u k1+k2+...+kr−1 k1!k2! . . . kr−1! = e ( (r−1)y ln c+ln a ln a+ln b ) u e (r−1) ( (r−1)x ln c+ln a ln a+ln b ) t ∑ k1+k2+...+kr−1≥0 Li(−k1,−k2,...,−kr−1)(1− e−2t) (1 + et)r−1 × ×u k1+k2+...+kr−1 k1!k2! . . . kr−1! = e ( (r−1)y ln c+ln a ln a+ln b ) u e (r−1) ( (r−1)x ln c+ln a ln a+ln b ) t (1 + et)r−1 ∑ 0