ON THE RIEMANN-LIOUVILLE FRACTIONAL q-INTEGRAL ...

19 downloads 0 Views 288KB Size Report
The present paper envisages the applications of Riemann-Liouville frac- tional q-integral operator to a basic analogue of Fox H-function. Results involving the ...
ON THE RIEMANN-LIOUVILLE FRACTIONAL q-INTEGRAL OPERATOR INVOLVING A BASIC ANALOGUE OF FOX H-FUNCTION S. L. Kalla∗ , R. K. Yadav∗∗ and S. D. Purohit∗∗

Abstract The present paper envisages the applications of Riemann-Liouville fractional q-integral operator to a basic analogue of Fox H-function. Results involving the basic hypergeometric functions like Gq (.), Jv (x; q), Yv (x; q), Kv (x; q), Hv (x; q) and various other q-elementary functions associated with the Riemann-Liouville fractional q-integral operator have been deduced as special cases of the main result. 2000 Mathematics Subject Classification: 33D60, 26A33, 33C60 Key Words and Phrases: fractional q-integral, H-function, q-functions 1. Introduction Agarwal [1] introduced the q-analogue of the Riemann-Liouville fractional integral operator as follows: Iqµ f (x)

1 = Γq (µ)

Zx (x − yq)µ−1 f (y)d(y; q), 0

where ‘µ‘ is an arbitrary order of integration such that Re(µ) > 0.

(1.1)

314

S.L. Kalla, R.K. Yadav, S.D. Purohit

Following Jackson [5], Al-Salam [2] and Agarwal [1], we have the basic integration defined as: Zx f (t)d(t; q) = x(1 − q)

∞ X

q k f (xq k ).

(1.2)

k=0

0

In view of equation (1.2), (1.1) can be expressed as: ∞

Iqµ f (x)

xµ (1 − q) X k = q (1 − q k+1 )µ−1 f (xq k ), Γq (µ)

(1.3)

k=0

where Re(µ) > 0. Further, for real or complex α and 0 < |q| < 1, the q-factorial is defined as: ½ 1 ; n=0 α (α; q)n ≡ (q ; q)n = α α+1 α+n−1 (1 − q )(1 − q )...(1 − q ) ; n = 1, 2, ... (1.4) or equivalently, (α; q)n =

∞ Y (α; q)∞ (1 − αq j ) = . n+j (1 − αq ) (αq n ; q)∞

(1.5)

j=0

In terms of the basic analogue of the gamma function, we have (α; q)n =

Γq (α + n)(1 − q)n , n > 0, Γq (α)

(1.6)

where the q-gamma function, cf. Gasper and Rahman [4], in various form is given by Γq (α) =

(q; q)∞ (1 − q)α−1 (q; q)α−1 = = , α−1 α−1 − q) (1 − q) (1 − q)α−1

(q α ; q)∞ (1

(1.7)

(α 6= 0, −1, −2, ...). Indeed

(q α ; q)n lt = (α)n , q → 1− (1 − q)n

(1.8)

(α)n = α(α + 1)...(α + n − 1).

(1.9)

where

ON THE RIEMANN-LIOUVILLE FRACTIONAL . . . The q-binomial series, cf. Gasper and Rahman [4], is given by   α; (αx; q)∞ q, x  = . 1 φ0  (x; q)∞ −;

315

(1.10)

Saxena, Modi and Kalla [10], introduced the basic analogue of the Fox Hfunction in the following manner: ¯ · ¸ ¯ (a, α) m1 ,n1 x; q ¯¯ HA,B (1.11) (b, β)

1 = 2π i

m Q1

Z

j=1 B Q

G(q 1−αj +αj s )πxs

j=1 A Q

G(q 1−bj +βj s )

C

n1 Q

s

G(q bj −βj )

j=m1 +1

ds, G(q αj −αj s )G(q 1−s ) sin πs

j=n1 +1

where 0 ≤ m1 ≤ B, 0 ≤ n1 ≤ A; αi0 s and βj0 s are all positive integers, the contour C is a line parallel to Re(ws), with indentations if necessary, in such a manner that all poles of G(q bj −βj s ), 1 ≤ j ≤ m1 are to the right, and those of G(q 1−αj +αj s ), 1 ≤ j ≤ n1 , to the left of C. The integral converges if Re[s log(x) − log sin πs] < 0 for large values of |s| on the contour, i.e. if |{arg(x) − w2 w1−1 log |x|}| < π, where 0 < |q| < 1, log q = −w = (w1 + iw2 ), w, w1 , w2 are definite quantities w1 and w2 being real. Also (∞ )−1 Y 1 α α+n G(q ) = (1 − q ) = α . (1.12) (q ; q)∞ n=0

If we set αj = βi = 1, 1 ≤ j ≤ A, 1 ≤ i ≤ B in (1.11), then it reduces to the basic Meijer’s G-function, namely ¯ ¯ ¸ · · ¸ ¯ (a, 1) ¯ a1 , ..., aA m1 ,n1 m1 n1 ¯ ¯ HA,B ≡ GA,B x; q ¯ x; q ¯ (b, 1) b1 , ..., bB

1 = 2π i

m Q1

Z

G(q bj −s )

j=1 B Q

C j=m1 +1

G(q 1−bj +s )

n1 Q

G(q 1−aj +s )πxs

j=1 A Q

ds, G(q αj −s )G(q 1−s ) sin πs

j=n1 +1

where 0 ≤ m1 ≤ B, 0 ≤ n1 ≤ A; and Re[s log(x) − log sin πs] < 0.

(1.13)

316

S.L. Kalla, R.K. Yadav, S.D. Purohit

Further, if we set n1 = 0, m1 = B in (1.13), we get the basic analogue of MacRobert’s E-function ¯ ¸ · ¯ a1 , ..., aA B,0 ¯ = Eq [B; bj : A; aj : x]. (1.14) GA,B x; q ¯ b1 , ..., bB Saxena and Kumar [11], introduced the basic analogues of Jv (x), Yv (x), Kv (x), Hv (x) in terms of Gq (.) function as follows: ¯ · 2 ¸ 2 ¯ − 2 1,0 x (1 − q) ¯ Jv (x; q) = {G(q)} G0,3 ; q ¯ v −v , (1.15) 4 2, 2 , 1 where Jv (x; q) denotes the q-analogue of Bessel function Jv (x). ¯ · 2 ¸ 2 −v−1 ¯ 2,0 x (1 − q) 2 2 ¯ Yv (x; q) = {G(q)} G1,4 ; q ¯ v −v −v−1 , 4 2, 2 , 2 , 1

(1.16)

where Yv (x; q) denotes the q-analogue of the Bessel function Yv (x). ¯ · 2 ¸ 2 ¯ − 2,0 x (1 − q) Kv (x; q) = (1 − q) G0,3 ; q ¯¯ v −v , (1.17) 4 2, 2 , 1 where Kv (x; q) denotes the basic analogue of the Bessel function of the third kind Kv (x). ¯ µ ¶ ¸ · 1+α 1 − q 1−α 3,1 x2 (1 − q)2 ¯¯ Hv (x; q) = ; q ¯ v −v 21+α , (1.18) G1,4 2 4 2, 2 , 2 , 1 where Hv (x; q) is the basic analogue of Struve’s function Hv (x). In view of the definition (1.11), the following elementary basic (q-) functions are expressible in terms of the basic analogue of Meijer’s G-function as: ¯ · ¸ ¯ − ¯ eq (−x) = G(q)G1,0 x(1 − q); q (1.19) 0,2 ¯ 0, 1 ¯ · 2 ¸ 2 ¯ z− √ 1,0 x (1 − q) −1/2 2 ¯ {G(q)} G0,3 ;q¯ 1 (1.20) sinq (x) = π(1 − q) 4 2 , 0, 1 ¯ · 2 ¸ 2 ¯ − √ 1,0 x (1 − q) −1/2 2 ¯ cosq (x) = π(1 − q) {G(q)} G0,3 ;q¯ (1.21) 0, 12 , 1 4 ¯ ¸ · 2 √ π x (1 − q)2 ¯¯ − 1,0 −1/2 2 (1.22) sinhq (x) = (1 − q) {G(q)} G0,3 − ;q¯ 1 i 4 2 , 0, 1

ON THE RIEMANN-LIOUVILLE FRACTIONAL . . .

317

¯ ¸ · 2 √ x (1 − q)2 ¯¯ − 1,0 −1/2 2 coshq (x) = π(1 − q) {G(q)} G0,3 − ;q¯ . (1.23) 0, 12 , 1 4 A detailed account of various functions expressible in terms of the Meijer G-function or Fox H-function can be found in research monographs due to Mathai and Saxena [8], [9]. A systematic and unified development of a new generalized fractional calculus, closely related to special functions of a rather general nature is given in [6], [7]. In a recent paper, Yadav and Purohit [12] have investigated some applications of Riemann-Liouville fractional q-integral operator to various basic hypergeometric functions of one variable. The motive for the present paper is to evaluate the Riemann-Liouville fractional basic integral operator involving basic analogue of the H-function and various other basic hypergeometric functions. The results deduced are believed to find certain applications to the solutions of basic (q-) integral equations. 2. Main results In this section, we shall evaluate the following q-fractional integral of the Riemann-Liouville type involving a basic analogue of Fox’s H-functions: ¯ ½ · ¸¾ ¯ (a, α) m1 ,n1 µ λ ¯ Iq HA,B ρx ; q ¯ (b, β) ¯ · ¸  ¯ (0, λ), (a, α)  m1 ,n1 +1 µ µ λ ¯  x (1 − q) HA+1,B+1 ρx ; q ¯ ,λ ≥ 0   (b, β), (−µ, λ)  (2.1) = ¯ · ¸   ¯ (a, α), (1 − λ)  m +1,n 1 1   xµ (1 − q)µ HA+1,B+1 ρxλ ; q ¯¯ ,λ < 0 (1 + µ, −λ), (b, β) where 0 ≤ m1 ≤ B, 0 ≤ n1 ≤ A. P r o o f o f (2.1). To prove (2.1), we apply equations (1.3) and (1.11) to the left hand side to obtain ∞ k µ X q (q ; q)k µ µ x (1 − q) (q; q)k k=0

1 × 2πi

m Q1

Z

j=1 B Q

C j=m1 +1

G(q bj −βj s )

n1 Q

G(q 1−aj +αj s )π(pxλ q λk )s

j=1

G(q 1−bj +βj s )

A Q j=n1 +1

ds. G(q aj −αj s )G(q 1−s ) sin πs

318

S.L. Kalla, R.K. Yadav, S.D. Purohit

Interchanging the order of summation and integration, which is valid under conditions given with (1.11), the above expression reduces to xµ (1

− 2πi

q)µ

m Q1

Z

G(q bj −βj s )

j=1 B Q j=m1 +1

×

G(q 1−aj +αj s )π(pxλ )s

j=1

G(q 1−bj +βj s )

C

n1 Q

A Q

G(q aj −αj s )G(q 1−s ) sin πs

j=n1 +1 ∞ k(1+λs) µ X q (q ; q)k k=0

(q; q)k

ds,

on summing inner 1 φ0 (.) series with the help of (1.10), it yields after some simplifications xµ (1 − q)µ 2πi m Q1

Z

j=1

×

B Q C j=m1 +1

G(q bj −βj s )

n1 Q

G(q 1−aj +αj s )G(q 1+λs )π(pxλ )s

j=1

G(q 1−bj +βj s )G(q 1+µ+λs )

A Q

ds, G(q aj −αj s )G(q 1−s ) sin πs

j=n1 +1

which on interpretation in view of (1.11), leads us to the right hand side of the result (2.1). The second part follows similarly. 3. Applications In this section, we shall derive certain basic integrals of the RiemannLiouville type, involving various basic functions expressible in terms of the basic analogue of Fox’s H-function or Meijer’s G-function, as the application of the main result (2.1). These results are presented in tabular form, see Table 1. Proofs of the results (3.1) and (3.2) follows directly from (2.1) with λ = ρ = 1 and on using the definitions (1.13)-(1.14) respectively. While if we assign m1 = 1, n1 = A = 0, B = 3, b1 = v/2, b2 = −v/2, b3 = 2 1, λ = 2, ρ = (1−q) in equation (2.1), we obtain 4 Iqµ

¯ ½ · 2 ¸¾ 2 ¯ − 1,0 x (1 − q) H0,3 = xµ (1 − q)µ ; q ¯¯ v ( 2 , 1), ( −v , 1), (1, 1) 4 2

ON THE RIEMANN-LIOUVILLE FRACTIONAL . . . · 1,1 × H1,4

¯ ¸ x2 (1 − q)2 ¯¯ (0, 2), ;q¯ v , ( 2 , 1), ( −v 4 2 , 1)(1, 1), (−µ, 2)

319

(3.12)

in view of of the definitions (1.15), the left hand side of (3.12) reduces to Iqµ {Jv (x; q)} = xµ (1 − q)µ {G(q)}2 · 1,1 × H1,4

¯ ¸ x2 (1 − q)2 ¯¯ (0, 2) ;q¯ v . ( 2 , 1), ( −v 4 2 , 1)(1, 1), (−µ, 2)

(3.13)

This completes the proof of (3.3). The proof of (3.4)-(3.6) follows similarly. To prove (3.7), we take m1 = 1, n1 = A = 0, B = 2, αi = βj = 1, b1 = 0, b2 = 1, λ = 1, ρ = (1 − q) in the main result (2.1) and then on making use of the definitions(1.13) and (1.19) we arrive at (3.7). The results (3.8)-(3.11) can be proved similarly by assigning particular values to the parameters m1 , n1 , A, B, λ and ρ, keeping in view the definitions (1.20)-(1.23). Finally, it is interesting to observe that in view of a limit formula for Gq (.) function due to Saxena and Kumar [11], if we let q → 1−1 , in (3.1), we obtain a known result mentioned in Erd´elyi [3] [eq. no. 96, table (13.1), p. 200]. Acknowledgment S. D. Purohit is grateful to CSIR, Govt. of INDIA, for providing a Senior Research Fellowship (No. 9/98(51) 2003. EMR.I) to enable him to carry out the present investigations.

320

S.L. Kalla, R.K. Yadav, S.D. Purohit

Table 1

ON THE RIEMANN-LIOUVILLE FRACTIONAL . . .

Table 1, cont’d

321

322

S.L. Kalla, R.K. Yadav, S.D. Purohit References

1. R. P. Agarwal, Certain fractional q-integrals and q-derivatives. Proc. Camb. Phil. Soc. 66 (1969), 365-370. 2. W. A. Al-Salam, Some fractional q-integrals and q-derivatives. Proc. Edin. Math. Soc. 15 (1966), 135-140. 3. A. Erd´elyi (Editor), Tables of Integral Transforms. Vol. II, McGrawHill Book Co. Inc., New York (1954). 4. G. Gasper, G. and M. Rahman, Basic Hypergeometric Series. Cambridge University Press, Cambridge (1990). 5. F. H. Jackson, Basic integration. Quart. J. Math. (Oxford), 2 (1951), 1-16. 6. S. L. Kalla and V. S. Kiryakova, An H-function generalized fractional calculus based upon composition of Erd´elyi-Kober operators in Lp. Math. Japonica 35 (1990), 1151-1171. 7. V. S. Kiryakova, Generalized Fractional Calculus and Applications. Longman Scientific & Tech., Essex (1994). 8. A. Mathai and R. K. Saxena, Generalized Hypergeometric Functions With Applications in Statistics and Physical Sciences. Springer-Verlag, Berlin (1973). 9. Mathai, A. M and Saxena, R. K.: The H-function With Application in Statistics and Other Disciplines. John Wiley & Sons. Inc. New York. (1978). 10. R. K. Saxena, G. C. Modi and S. L. Kalla, A basic analogue of Fox’s H-function. Rev. Tec. Ing., Univ. Zulia 6 (1983), 139-143. 11. R. K. Saxena and R. Kumar, Recurrence relations for the basic analogue of the H-function. J. Nat. Acad. Math. 8 (1990), 48-54. 12. R. K. Yadav and S. D. Purohit, Application of Riemann-Liouville fractional q-integral operator to basic hypergeometric functions. To appear in: Acta Ciencia Indica 30, No. 3 (2004). ∗

Department of Mathematics and Computer Science P.O. Box-5969, Safat 13060, KUWAIT e-mail: [email protected] ∗∗

Department of Mathematics & Statistics J. N. Vyas University, Jodhpur-342001 (INDIA) e-mail: [email protected]

Received: 24 April 2005