Abstract

By using fermionic p-adic q-integral on 𝑝, we give some interesting relationship between the twisted (h, q)-Euler numbers with weight α and the q-Bernstein polynomials.

1. Introduction

Let 𝑝 be a fixed odd prime number. Throughout this paper, we always make use of the following notations: denotes the ring of rational integers, 𝑝 denotes the ring of 𝑝-adic rational integers, 𝑝 denotes the field of 𝑝-adic rational numbers, and 𝑝 denotes the completion of algebraic closure of 𝑝, respectively. Let be the set of natural numbers and +={0}. Let 𝐶𝑝𝑛={𝑤𝑤𝑝𝑛=1} be the cyclic group of order 𝑝𝑛, and let 𝑇𝑝=lim𝑛𝐶𝑝𝑛=𝐶𝑝=𝑛0𝐶𝑝𝑛,(1.1) (see [122]), be the locally constant space. For 𝑤𝑇𝑝, we denote by 𝜙𝑤𝑝𝑝 the locally constant function 𝑥𝑤𝑥. The 𝑝-adic absolute value is defined by |𝑥|𝑝=1/𝑝𝑟, where 𝑥=𝑝𝑟𝑠/𝑡(𝑟and𝑠,𝑡with(𝑠,𝑡)=(𝑝,𝑠)=(𝑝,𝑡)=1). In this paper, we assume that 𝑞𝑝 with |𝑞1|𝑝<1 as an indeterminate. The 𝑞-number is defined by [𝑥]𝑞=1𝑞𝑥1𝑞,(1.2) (see [122]). Note that lim𝑞1[𝑥]𝑞=𝑥. For 𝑓𝑈𝐷𝑝=𝑓𝑓𝑝𝑝isuniformlydierentiablefunction,(1.3) the fermionic 𝑝-adic 𝑞-integral on 𝑝 is defined by Kim as follows: 𝐼𝑞(𝑓)=𝑝𝑓(𝑥)𝑑𝜇𝑞(𝑥)=lim𝑁1+𝑞1+𝑞𝑝𝑁𝑝𝑁1𝑥=0𝑓(𝑥)(𝑞)𝑥,(1.4) (see [17]). From (1.4), we note that 𝑞𝑛𝐼𝑞𝑓𝑛=(1)𝑛𝐼𝑞(𝑓)+[2]𝑞𝑛1𝑙=0(1)𝑛1𝑙𝑞𝑙𝑓(𝑙),(1.5) where 𝑓𝑛(𝑥)=𝑓(𝑥+𝑛) for 𝑛.

For 𝑘,𝑛+ and 𝑥[0,1], Kim defined 𝑞-Bernstein polynomials, which are different 𝑞-Bernstein polynomials of Phillips, as follows: 𝐵𝑘,𝑛(𝑥,𝑞)=𝑛𝑘[𝑥]𝑘𝑞[1𝑥]𝑛𝑘𝑞1,(1.6) (see [5]). In [9], the 𝑝-adic extension of (1.6) is given by 𝐵𝑘,𝑛(𝑥,𝑞)=𝑛𝑘[𝑥]𝑘𝑞[1𝑥]𝑛𝑘𝑞1,where𝑥𝑝,𝑛,𝑘+.(1.7) For 𝛼, , 𝑤𝑇𝑝, and 𝑞𝑝 with |1𝑞|𝑝1, twisted (,𝑞)-Euler numbers 𝐸(,𝛼)𝑛,𝑞,𝑤 with weight 𝛼 are defined by 𝐸(,𝛼)𝑛,𝑞,𝑤=𝑝𝜙𝑤(𝑥)𝑞𝑥(1)[𝑥]𝑛𝑞𝛼𝑑𝜇𝑞(𝑥).(1.8) In the special case, 𝑥=0, 𝐸(,𝛼)𝑛,𝑞,𝑤(0)=𝐸(,𝛼)𝑛,𝑞,𝑤 are called the 𝑛-th twisted (,𝑞)-Euler numbers with weight 𝛼.

In this paper, we investigate some relations between the 𝑞-Bernstein polynomials and the twisted (,𝑞)-Euler numbers with weight 𝛼. From these relations, we derive some interesting identities on the twisted (,𝑞)-Euler numbers and polynomials with weight 𝛼.

2. Twisted (,𝑞)-Euler Numbers and Polynomials with Weight 𝛼

By using 𝑝-adic 𝑞-integral and (1.8), we obtain 𝑝𝜙𝑤(𝑥)𝑞𝑥(1)[𝑥]𝑛𝑞𝛼𝑑𝜇𝑞(𝑥)=lim𝑁1𝑝𝑁𝑞𝑝𝑁1𝑥=0[𝑥]𝑛𝑞𝛼𝑤𝑥𝑞𝑥(1)(𝑞)𝑥=[2]𝑞11𝑞𝛼𝑛𝑛𝑙=0𝑛𝑙(1)𝑙11+𝑤𝑞𝛼𝑙+.(2.1) We set 𝐹(,𝛼)𝑞,𝑤(𝑡)=𝑛=0𝐸(,𝛼)𝑛,𝑞,𝑤𝑡𝑛𝑛!.(2.2) By (2.1) and (2.2), we have 𝐹(,𝛼)𝑞,𝑤(𝑡)=𝑛=0𝐸(,𝛼)𝑛,𝑞,𝑤𝑡𝑛𝑛!=[2]𝑞𝑛=011𝑞𝛼𝑛𝑛𝑙=0𝑛𝑙(1)𝑙11+𝑤𝑞𝛼𝑙+𝑡𝑛𝑛!=[2]𝑞𝑚=0(1)𝑚𝑤𝑚𝑞𝑚𝑒[𝑚]𝑞𝛼𝑡.(2.3) Since [𝑥+𝑦]𝑞𝛼=[𝑥]𝑞𝛼+𝑞𝛼𝑥[𝑦]𝑞𝛼, we obtain 𝐸(,𝛼)𝑛,𝑞,𝑤(𝑥)=𝑝𝜙𝑤(𝑦)𝑞𝑦(1)[𝑦+𝑥]𝑞𝛼𝑡𝑑𝜇𝑞(𝑦)=𝑛𝑙=0𝑛𝑙𝑞𝛼𝑥𝑙[𝑥]𝑛𝑙𝑞𝛼𝑝𝜙𝑤(𝑦)𝑞𝑦(1)[𝑦]𝑙𝑞𝛼𝑑𝜇𝑞(𝑦)=𝑛𝑙=0𝑛𝑙𝑞𝛼𝑥𝑙[𝑥]𝑛𝑙𝑞𝛼𝐸(,𝛼)𝑙,𝑞,𝑤.(2.4)

Therefore, we obtain the following theorem.

Theorem 2.1. For 𝑛+ and 𝑤𝑇𝑝, we have 𝐸(,𝛼)𝑛,𝑞,𝑤(𝑥)=[2]𝑞𝑚=0(1)𝑚𝑤𝑚𝑞𝑚[𝑥+𝑚]𝑛𝑞𝛼.(2.5) Furthermore, 𝐸(,𝛼)𝑛,𝑞,𝑤(𝑥)=𝑛𝑙=0𝑛𝑙𝑞𝛼𝑥𝑙[𝑥]𝑛𝑙𝑞𝛼𝐸(,𝛼)𝑙,𝑞,𝑤=[𝑥]𝑞𝛼+𝑞𝛼𝑥𝐸(,𝛼)𝑞,𝑤𝑛,(2.6) with usual convention about replacing (𝐸(,𝛼)𝑞,𝑤)𝑛 with 𝐸(,𝛼)𝑛,𝑞,𝑤.
Let 𝐹(,𝛼)𝑞,𝑤(𝑡,𝑥)=𝑛=0𝐸(,𝛼)𝑛,𝑞,𝑤(𝑥)𝑡𝑛/𝑛!. Then we see that 𝐹(,𝛼)𝑞,𝑤(𝑡,𝑥)=[2]𝑞𝑚=0(1)𝑚𝑤𝑚𝑞𝑚𝑒[𝑥+𝑚]𝑞𝛼𝑡.(2.7) In the special case, 𝑥=0, let 𝐹(,𝛼)𝑞,𝑤(𝑡,0)=𝐹(,𝛼)𝑞,𝑤(𝑡).
By (2.1), we get 𝐸(,𝛼)𝑛,𝑞1,𝑤1(1𝑥)=(1)𝑛𝑤𝑞𝛼𝑛+1𝐸(,𝛼)𝑛,𝑞,𝑤(𝑥).(2.8) From (2.3) and (2.7), we note that 𝑤𝑞𝐹(,𝛼)𝑞,𝑤(𝑡,1)+𝐹(,𝛼)𝑞,𝑤(𝑡)=[2]𝑞.(2.9) By (2.9), we get the following recurrence formula: 𝐸(,𝛼)0,𝑞,𝑤=[2]𝑞1+𝑞𝑤,𝑞𝑤𝐸(,𝛼)𝑛,𝑞,𝑤(1)+𝐸(,𝛼)𝑛,𝑞,𝑤=0if𝑛>0.(2.10)

By (2.10) and Theorem 2.1, we obtain the following theorem.

Theorem 2.2. For 𝑛+ and 𝑤𝑇𝑝, we have 𝐸(,𝛼)0,𝑞,𝑤=[2]𝑞1+𝑞𝑤,𝑞𝑤𝑞𝛼𝐸(,𝛼)𝑞,𝑤+1𝑛+𝐸(,𝛼)𝑛,𝑞,𝑤=0if𝑛>0,(2.11) with usual convention about replacing (𝐸(,𝛼)𝑞,𝑤)𝑛 with 𝐸(,𝛼)𝑛,𝑞,𝑤.
By (2.4), Theorem 2.1, and Theorem 2.2, we have 𝑞2𝑤2𝐸(,𝛼)𝑛,𝑞,𝑤(2)[2]𝑞1+𝑞𝑤𝑞2𝑤2[2]𝑞1+𝑞𝑤𝑞𝑤=𝑞2𝑤2𝑛𝑙=0𝑛𝑙𝑞𝛼𝑙𝑞𝛼𝐸(,𝛼)𝑞,𝑤+1𝑙[2]𝑞1+𝑞𝑤𝑞2𝑤2[2]𝑞1+𝑞𝑤𝑞𝑤=𝑞2𝑤2𝑛𝑙=1𝑛𝑙𝑞𝛼𝑙𝑞𝛼𝐸(,𝛼)𝑞,𝑤+1𝑙[2]𝑞1+𝑞𝑤𝑞𝑤=𝑞𝑤𝑛𝑙=1𝑛𝑙𝑞𝛼𝑙𝐸(,𝛼)𝑙,𝑞,𝑤[2]𝑞1+𝑞𝑤𝑞𝑤=𝑞𝑤𝑛𝑙=0𝑛𝑙𝑞𝛼𝑙𝐸(,𝛼)𝑙,𝑞,𝑤=𝑞𝑤𝐸(,𝛼)𝑛,𝑞,𝑤(1)=𝐸(,𝛼)𝑛,𝑞,𝑤if𝑛>0.(2.12)

Therefore, we obtain the following theorem.

Theorem 2.3. For 𝑛, we have 𝐸(,𝛼)𝑛,𝑞,𝑤(2)=1𝑞2𝑤2𝐸(,𝛼)𝑛,𝑞,𝑤+[2]𝑞1+𝑞𝑤+1𝑞𝑤[2]𝑞1+𝑞𝑤.(2.13)
By (2.8), we see that 𝑞1𝑤𝑝[1𝑥]𝑛𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=(1)𝑛𝑞𝛼𝑛+1𝑤𝑝[𝑥1]𝑛𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=(1)𝑛𝑞𝛼𝑛+1𝑤𝐸(,𝛼)𝑛,𝑞,𝑤(1)=𝐸(,𝛼)𝑛,𝑞1,𝑤1(2).(2.14)

Therefore, we obtain the following theorem.

Theorem 2.4. For 𝑛+, we have 𝑞1𝑤𝑝[1𝑥]𝑛𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝐸(,𝛼)𝑛,𝑞1,𝑤1(2).(2.15) Let 𝑛. By Theorems 2.3 and 2.4, we get 𝑞1𝑤𝑝[1𝑥]𝑛𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑞2𝑤2𝐸(,𝛼)𝑛,𝑞1,𝑤1+𝑞1𝑤[2]𝑞1+𝑞𝑤+𝑞21𝑤2[2]𝑞1+𝑞𝑤.(2.16)
From (2.16), we have 𝑝[1𝑥]𝑛𝑞1𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑞+1𝑤𝐸(,𝛼)𝑛,𝑞1,𝑤1+[2]𝑞1+𝑞𝑤+𝑞𝑤[2]𝑞1+𝑞𝑤.(2.17)

Therefore, we obtain the following corollary.

Corollary 2.5. For 𝑛, we have 𝑝[1𝑥]𝑛𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑞+1𝑤𝐸(,𝛼)𝑛,𝑞1,𝑤1+[2]𝑞.(2.18)
Kim defined the 𝑞-Bernstein polynomials with weight 𝛼 of degree 𝑛 as below.
For 𝑥𝑝, the 𝑝-adic 𝑞-Bernstein polynomials with weight 𝛼 of degree 𝑛 are given by 𝐵(𝛼)𝑘,𝑛(𝑥,𝑞)=𝑛𝑘[𝑥]𝑘𝑞𝛼[1𝑥]𝑛𝑘𝑞𝛼,where𝑛,𝑘+.(2.19) compare [5, 10, 22] By (2.19), we get the symmetry of 𝑞-Bernstein polynomials as follows: 𝐵(𝛼)𝑘,𝑛(𝑥,𝑞)=𝐵(𝛼)𝑛𝑘,𝑛1𝑥,𝑞1,(2.20) see [8]. Thus, by Corollary 2.5, (2.19), and (2.20), we see that 𝑝𝐵(𝛼)𝑘,𝑛(𝑥,𝑞)𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑝𝐵(𝛼)𝑛𝑘,𝑛1𝑥,𝑞1𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛𝑘𝑘𝑙=0𝑘𝑙(1)𝑘+𝑙𝑝[1𝑥]𝑛𝑙𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛𝑘𝑘𝑙=0𝑘𝑙(1)𝑘+𝑙𝑞+1𝑤𝐸(,𝛼)𝑛𝑙,𝑞1,𝑤1+[2]𝑞.(2.21)
For 𝑛,𝑘+ with 𝑛>𝑘, we have 𝑝𝐵(𝛼)𝑘,𝑛(𝑥,𝑞)𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛𝑘𝑘𝑙=0𝑘𝑙(1)𝑘+𝑙𝑞+1𝑤𝐸(,𝛼)𝑛𝑙,𝑞1,𝑤1+[2]𝑞=𝑞+1𝑤𝐸(,𝛼)𝑛,𝑞1,𝑤1+[2]𝑞if𝑘=0,𝑞+1𝑤𝑛𝑘𝑘𝑙=0𝑘𝑙(1)𝑘+𝑙𝐸(,𝛼)𝑛𝑙,𝑞1,𝑤1if𝑘>0.(2.22) Let us take the fermionic 𝑞-integral on 𝑝 for the 𝑞-Bernstein polynomials with weight 𝛼 of degree 𝑛 as follows: 𝑝𝐵(𝛼)𝑘,𝑛(𝑥,𝑞)𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛𝑘𝑝[𝑥]𝑘𝑞[1𝑥]𝑛𝑘𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛𝑘𝑛𝑘𝑙=0𝑛𝑘𝑙(1)𝑙𝐸(,𝛼)𝑙+𝑘,𝑞,𝑤.(2.23)

Therefore, by (2.22) and (2.23), we obtain the following theorem.

Theorem 2.6. Let 𝑛,𝑘+ with 𝑛>𝑘. Then we have 𝑝𝐵(𝛼)𝑘,𝑛(𝑥,𝑞)𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑞+1𝑤𝐸(,𝛼)𝑛,𝑞1,𝑤1+[2]𝑞if𝑘=0,𝑞+1𝑤𝑛𝑘𝑘𝑙=0𝑘𝑙(1)𝑘+𝑙𝐸(,𝛼)𝑛𝑙,𝑞1,𝑤1if𝑘>0.(2.24) Moreover, 𝑛𝑘𝑙=0𝑛𝑘𝑙(1)𝑙E(,𝛼)𝑙+𝑘,𝑞,𝑤=𝑞+1𝑤𝐸(,𝛼)𝑛,𝑞1,𝑤1+[2]𝑞if𝑘=0,𝑞+1𝑤𝑘𝑙=0𝑘𝑙(1)𝑘+𝑙𝐸(,𝛼)𝑛𝑙,𝑞1,𝑤1if𝑘>0.(2.25)
Let 𝑛1,𝑛2,𝑘+ with 𝑛1+𝑛2>2𝑘. Then we get 𝑝𝐵(𝛼)𝑘,𝑛1(𝑥,𝑞)𝐵(𝛼)𝑘,𝑛2(𝑥,𝑞)𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛2𝑘2𝑘𝑙=02𝑘𝑙(1)𝑙+2𝑘𝑝[1𝑥]𝑛1+𝑛2𝑙𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛2𝑘2𝑘𝑙=02𝑘𝑙(1)𝑙+2𝑘𝑞+1𝑤𝐸(,𝛼)𝑛1+𝑛2𝑙,𝑞1,𝑤1+[2]𝑞=𝑞+1𝑤𝐸(,𝛼)𝑛1+𝑛2,𝑞1,𝑤1+[2]𝑞if𝑘=0,𝑛1𝑘𝑛2𝑘2𝑘𝑙=02𝑘𝑙(1)2𝑘+𝑙𝑞+1𝑤𝐸(,𝛼)𝑛1+𝑛2𝑙,𝑞1,𝑤1+[2]𝑞if𝑘0.(2.26)

Therefore, by (2.26), we obtain the following theorem.

Theorem 2.7. For 𝑛1,𝑛2,𝑘+ with 𝑛1+𝑛2>2𝑘, we have 𝑝𝐵(𝛼)𝑘,𝑛1(𝑥,𝑞)𝐵(𝛼)𝑘,𝑛2(𝑥,𝑞)𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑞+1𝑤𝐸(,𝛼)𝑛1+𝑛2,𝑞1,𝑤1+[2]𝑞if𝑘=0,𝑞+1𝑤𝑛1𝑘𝑛2𝑘2𝑘𝑙=02𝑘𝑙(1)2𝑘+𝑙𝐸(,𝛼)𝑛1+𝑛2𝑙,𝑞1,𝑤1if𝑘0.(2.27)
From the binomial theorem, we can derive the following equation: 𝑝𝐵(𝛼)𝑘,𝑛1(𝑥,𝑞)𝐵(𝛼)𝑘,𝑛2(𝑥,𝑞)𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛2𝑘𝑛1+𝑛22𝑘𝑙=0(1)𝑙𝑛1+𝑛22𝑘𝑙𝑝[𝑥]2𝑘+𝑙𝑞𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛2𝑘𝑛1+𝑛22𝑘𝑙=0(1)𝑙𝑛1+𝑛22𝑘𝑙𝐸(,𝛼)2𝑘+𝑙,𝑞,𝑤.(2.28)

Thus, by (2.28) and Theorem 2.7, we obtain the following corollary.

Corollary 2.8. Let 𝑛1,𝑛2,𝑘+ with 𝑛1+𝑛2>2𝑘. Then we have 𝑛1+𝑛22𝑘𝑙=0(1)𝑙𝑛1+𝑛22𝑘𝑙𝐸(,𝛼)2𝑘+𝑙,𝑞=𝑞+1𝑤𝐸(,𝛼)𝑛1+𝑛2,𝑞𝛼,𝑤1+[2]𝑞if𝑘=0,𝑞+1𝑤2𝑘𝑙=02𝑘𝑙(1)2𝑘+𝑙𝐸(,𝛼)𝑛1+𝑛2𝑙,𝑞𝛼,𝑤1if𝑘>0.(2.29)
For 𝑥𝑝 and 𝑠 with 𝑠2, let 𝑛1,𝑛2,,𝑛𝑠,𝑘+ with 𝑛1++𝑛𝑠>𝑠𝑘. Then we take the fermionic 𝑝-adic 𝑞-integral on 𝑝 for the 𝑞-Bernstein polynomials with weight 𝛼 of degree 𝑛 as follows: 𝑝𝐵(𝛼)𝑘,𝑛1(𝑥,𝑞)𝐵(𝛼)𝑘,𝑛𝑠(𝑥,𝑞)𝑠-times𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛𝑠𝑘𝑝[𝑥]𝑠𝑘𝑞[1𝑥]𝑛1++𝑛𝑠𝑠𝑘𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛s𝑘𝑠𝑘𝑙=0𝑠𝑘𝑙(1)𝑙+𝑠𝑘𝑝[1𝑥]𝑛1++𝑛𝑠𝑙𝑞𝛼𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛𝑠𝑘𝑠𝑘𝑙=0𝑠𝑘𝑙(1)𝑙+𝑠𝑘𝑞+1𝑤𝐸(,𝛼)𝑛1++𝑛𝑠𝑙,𝑞1,𝑤1+[2]𝑞=𝑞+1𝑤𝐸(,𝛼)𝑛1++𝑛𝑠,𝑞1,𝑤1+[2]𝑞if𝑘=0,𝑞+1𝑤𝑛1𝑘𝑛𝑠𝑘𝑠𝑘𝑙=0𝑠𝑘𝑙(1)𝑙+𝑠𝑘𝐸(,𝛼)𝑛1++𝑛𝑠𝑙,𝑞1,𝑤1if𝑘>0.(2.30)

Therefore, by (2.30), we obtain the following theorem.

Theorem 2.9. For 𝑠 with 𝑠2, let 𝑛1,𝑛2,,𝑛𝑠,𝑘+ with 𝑛1++𝑛𝑠>𝑠𝑘. Then we get 𝑝𝐵(𝛼)𝑘,𝑛1(𝑥,𝑞)𝐵(𝛼)𝑘,𝑛𝑠(𝑥,𝑞)𝑠-times𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑞+1𝑤𝐸(,𝛼)𝑛1++𝑛𝑠,𝑞1,𝑤1+[2]𝑞if𝑘=0,𝑞+1𝑤𝑛1𝑘𝑛𝑠𝑘𝑠𝑘𝑙=0𝑠𝑘𝑙(1)𝑙+𝑠𝑘𝐸(,𝛼)𝑛1++𝑛𝑠𝑙,𝑞1,𝑤1if𝑘>0.(2.31)
By the definition of 𝑞-Bernstein polynomials with weight 𝛼 and the binomial theorem, we easily get 𝑝𝐵(𝛼)𝑘,𝑛1(𝑥,𝑞)𝐵(𝛼)𝑘,𝑛𝑠(𝑥,𝑞)𝑠-times𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛𝑠𝑘𝑛1++𝑛𝑠𝑠𝑘𝑙=0(1)𝑙𝑛1++𝑛𝑠𝑠𝑘𝑙𝑝[𝑥]𝑠𝑘+𝑙𝑞𝑞(1)𝑥𝑤𝑥𝑑𝜇𝑞(𝑥)=𝑛1𝑘𝑛𝑠𝑘𝑛1++𝑛𝑠𝑠𝑘𝑙=0(1)𝑙𝑛1++𝑛𝑠𝑠𝑘𝑙𝐸(,𝛼)𝑠𝑘+𝑙,𝑞,𝑤.(2.32)

Therefore, we have the following corollary.

Corollary 2.10. For 𝑤𝑇𝑝,𝑠 with 𝑠2, let 𝑛1,𝑛2,,𝑛𝑠,𝑘+ with 𝑛1++𝑛𝑠>𝑠𝑘. Then we have 𝑛1++n𝑠𝑠𝑘𝑙=0(1)𝑙𝑛1++𝑛𝑠𝑠𝑘𝑙𝐸(,𝛼)𝑠𝑘+𝑙,𝑞,𝑤=𝑞+1𝑤𝐸(,𝛼)𝑛1++𝑛𝑠,𝑞1,𝑤1+[2]𝑞if𝑘=0,𝑞+1𝑤𝑠𝑘𝑙=0𝑠𝑘𝑙(1)𝑙+𝑠𝑘𝐸(,𝛼)𝑛1++𝑛𝑠𝑙,𝑞1,𝑤1if𝑘>0.(2.33)