We study a class of column upper-minus-lower (CUML) Toeplitz matrices, which are "close" to the Toeplitz matrices in the sense that their (1,−1)-cyclic displacements coincide with φ-cyclic displacement of some Toeplitz matrices. Among others, we derive the inverse formula for CUML Toeplitz matrices in the form of sums of products of factor circulants by constructing the corresponding displacement of the matrices. In addition, by the relationship between CUML Toeplitz matrices and CUML Hankel matrices, the inverse formula for CUML Hankel matrices is also obtained.
Citation: Yanpeng Zheng, Xiaoyu Jiang. Quasi-cyclic displacement and inversion decomposition of a quasi-Toeplitz matrix[J]. AIMS Mathematics, 2022, 7(7): 11647-11662. doi: 10.3934/math.2022649
[1] | Yijia Zhang, Tao Xie, Yunlong Ma . Robustness analysis of exponential stability of Cohen-Grossberg neural network with neutral terms. AIMS Mathematics, 2025, 10(3): 4938-4954. doi: 10.3934/math.2025226 |
[2] | R. Sriraman, R. Samidurai, V. C. Amritha, G. Rachakit, Prasanalakshmi Balaji . System decomposition-based stability criteria for Takagi-Sugeno fuzzy uncertain stochastic delayed neural networks in quaternion field. AIMS Mathematics, 2023, 8(5): 11589-11616. doi: 10.3934/math.2023587 |
[3] | Jingyang Ran, Tiecheng Zhang . Fixed-time synchronization control of fuzzy inertial neural networks with mismatched parameters and structures. AIMS Mathematics, 2024, 9(11): 31721-31739. doi: 10.3934/math.20241525 |
[4] | Wenxiang Fang, Tao Xie, Biwen Li . Robustness analysis of fuzzy BAM cellular neural network with time-varying delays and stochastic disturbances. AIMS Mathematics, 2023, 8(4): 9365-9384. doi: 10.3934/math.2023471 |
[5] | Wenxiang Fang, Tao Xie . Robustness analysis of stability of Takagi-Sugeno type fuzzy neural network. AIMS Mathematics, 2023, 8(12): 31118-31140. doi: 10.3934/math.20231593 |
[6] | Tao Xie, Wenqing Zheng . Robustness analysis of Cohen-Grossberg neural network with piecewise constant argument and stochastic disturbances. AIMS Mathematics, 2024, 9(2): 3097-3125. doi: 10.3934/math.2024151 |
[7] | Tiecheng Zhang, Wei He . Global robust stability of fuzzy cellular neural networks with parameter uncertainties. AIMS Mathematics, 2024, 9(4): 8063-8078. doi: 10.3934/math.2024392 |
[8] | Ankit Kumar, Subir Das, Vijay K. Yadav, Rajeev, Jinde Cao, Chuangxia Huang . Synchronizations of fuzzy cellular neural networks with proportional time-delay. AIMS Mathematics, 2021, 6(10): 10620-10641. doi: 10.3934/math.2021617 |
[9] | Chengqiang Wang, Xiangqing Zhao, Yang Wang . Finite-time stochastic synchronization of fuzzy bi-directional associative memory neural networks with Markovian switching and mixed time delays via intermittent quantized control. AIMS Mathematics, 2023, 8(2): 4098-4125. doi: 10.3934/math.2023204 |
[10] | Shuo Ma, Jiangman Li, Qiang Li, Ruonan Liu . Adaptive exponential synchronization of impulsive coupled neutral stochastic neural networks with Lévy noise and probabilistic delays under non-Lipschitz conditions. AIMS Mathematics, 2024, 9(9): 24912-24933. doi: 10.3934/math.20241214 |
We study a class of column upper-minus-lower (CUML) Toeplitz matrices, which are "close" to the Toeplitz matrices in the sense that their (1,−1)-cyclic displacements coincide with φ-cyclic displacement of some Toeplitz matrices. Among others, we derive the inverse formula for CUML Toeplitz matrices in the form of sums of products of factor circulants by constructing the corresponding displacement of the matrices. In addition, by the relationship between CUML Toeplitz matrices and CUML Hankel matrices, the inverse formula for CUML Hankel matrices is also obtained.
As a generalization of cyclic codes and negacyclic codes, constacyclic codes were first introduced by Berlekamp in 1968 [3]. Given a nonzero element λ in a finite filed Fq, a linear code C of length n over Fq is called λ-constacyclic if (λcn−1,c0,⋯,cn−2)∈C for every (c0,c1,⋯,cn−1)∈C. Constacyclic codes over finite fields form a remarkable class of linear codes, as it includes the class of cyclic codes and the class of negacyclic codes as proper subclasses. Constacyclic codes have rich algebraic structure so that they can be efficiently encoded and decoded by means of shift registers. Repeated-root constacyclic codes were a special class of constacyclic codes. Repeated-root constacyclic codes were first studied by Castagnoli et al. [4] and van Lint [13], and they showed that repeated-root cyclic codes have a concatenated construction and are not asymptotically good.
Recently, repeated-root constacyclic codes have been studied by many authors. To determine the generator polynomials of all constacyclic codes of arbitrary length over finite fields is an important problem. Dinh studied repeated-root constacyclic codes of lengths 2ps, 3ps, 4ps and 6ps in a series of papers [8,9,10,11]. He determined the algebraic structure of these repeated-root constacyclic codes over finite fields in terms of their generator polynomials. In [7], Chen et al. introduced an equivalence relation called isometry for the nonzero elements of Fq to classify constacyclic codes of length n over Fq. They have the same distance structures and the same algebraic structures for belonging to the same equivalence classes induced by isometry. Furthermore, in [5], Chen et al. considered a more specified relationship than isometry that enabled us to obtain more explicit description of generator polynomials of all constacyclic codes. According to the equivalence classes, all constacyclic codes of length ℓps over Fqm and their dual are characterized, where ℓ is a prime different from p and s is a positive integer. In 2012, Bakshi and Raka [1] also determined all Λ-constacyclic codes of length 2tps(t≥1,s≥0 are integers) over Fpr using different methods from Chen et al.. In 2015, Chen et al. [6] determined the algebraic structure of all constacyclic codes of length 2ℓmps over Fpr and their dual codes in terms of their generator polynomials, where ℓ,p are distinct odd primes and s,m are positive integers. In the conclusion of the paper [6], they proposed an open problem to study all constacyclic codes of length kℓmps over Fq, where p is the characteristic of Fq, ℓ is an odd prime different from p, and k is a prime different from ℓ and p. Batoul et al. [2] investigated the structure of constacyclic codes of length 2ampr over Fps with a≥1 and (m,p)=1. They also provided certain sufficient conditions under which these codes are equivalent to cyclic codes of length 2ampr over Fps. Sharma [16] determined all constacyclic codes of length ℓtps over Fpr and their dual codes, where ℓ,p are distinct primes, ℓ is odd and s,t,r are positive integers. In 2016, Sharma et al. [17] determine generator polynomials of all constacyclic codes of length 4ℓmpn over the finite field Fq and their dual codes, where p,ℓ are distinct odd primes, q is a power of p and m,n are positive integers. Working in the same direction, Liu et al. obtained generator polynomials of all repeated-root constacyclic codes of length 3ℓps over Fq in [14], where ℓ is an odd prime different from p and 3. In 2017, Liu et al. [15] explicitly determine the generator polynomials of all repeated-root constacyclic codes of length nℓps over Fq and their dual codes, where ℓ is an odd prime different from p, and n is an odd prime different from both ℓ and p such that n=2h+1 for some prime h. In 2019, Wu and Yue et al. [19,20] explicitly factorize the polynomial xn−λ for each λ∈Fq. As applications, they obtain all repeated-root λ-constacyclic codes and their dual codes of length nps over Fq.
In this paper, we answer the question of B. Chen, H. Dinh and Liu. That is we determine all the constacyclic codes of length p1pt2ps over Fq, where p is the characteristic of Fq, p1 is an odd prime different from p, and p1 is a prime different from p2 and p. We give the explicit generator polynomials of all the constacyclic codes of length p1pt2ps over Fq and their dual codes, and determine all self-dual cyclic codes of length p1pt2ps and their enumeration.
The remainder of this paper is organized as follows. In Section 2 we give a brief background on some basic results which we need in the following parts. In Section 3, we calculate the q-cyclotomic cosets modulo p1pt2 as a preparation for giving the generator polynomials of constacyclic codes of length p1pt2ps over Fq. In Section 4, we first describe a general method to obtain the generator polynomials of constacyclic codes, and then with this method and the results of q-cyclotomic cosets modulo p1pt2 we give the explicit generator polynomials of all the constacyclic codes of length p1pt2ps. And in Section 5, all the self-dual cyclic codes of length p1pt2ps over Fq are given. In the last section, as an example we calculate the case of length 5ℓps, where ℓ is a prime different from 5 and p.
In this section, we first review some basic results in number theory and finite fields, which we will in the following parts, and then give a brief introduction to the λ-constacyclic codes. For a positive integer n, we denote by Zn the ring of integers module n throughout this paper. Let p be a prime number, and q be a power of p. We denote by Fq the finite field with q elements, and fix a generator element ξ of the multiplicative group F∗q, that is, F∗q=⟨ξ⟩. In this paper, we mainly deal with the repeated-root constacyclic codes of length p1pt2ps over Fq, where p1 and p2 are two distinct odd prime numbers different from p. For any positive integer d and i=1,2, we write fi,d=ordpdi(q) for the multiplicative order of q modulo pdi, and set gi,d=ϕ(pdi)fi,d, where ϕ is the Euler's phi function. When d=1, we write fi=fi,1 and gi=gi,1 for simplicity. For i=1,2, there are positive integers ui and wi such that qfi=1+puiiwi and pi∤wi. Following the lifting-the-exponent lemma, we immediately have
fi,d=fipmax{0,d−ui}i. |
Lemma 2.1. [12] Assume that r is a primitive root of the odd prime p and (r+tp)p−1 is not congruent to 1 modulo p2. Then r+tp is a primitive root of pk for each k≥1.
Lemma 2.2. [18] Let n≥2 be an integer, and λ be a nonzero element in Fq with multiplicative order k=ord(λ). The binomial xn−λ is irreducible over Fq if and only if
(1) Every prime divisor of n divides k, but not q−1k;
(2) If 4∣n, then 4∣(q−1).
Let λ be a nonzero element in Fq. A λ-constacyclic code of length n is a linear code C such that (c0,c1,⋯,cn−1)∈C implies (λcn−1,c0,⋯,cn−2)∈C. This definition is a natural generalization of cyclic code and negacyclic code. A λ-constacyclic code C of length n over Fq can be regarded as an ideal (g(x)) of the quotient ring Fq[x]/(xn−λ), where g(x) is a divisor of xn−λ. Let C be a λ-constacyclic code of length n over Fq, then the dual code of code C is given by C⊥={x∈Fnq:x⋅y=0,∀y∈C}, where x⋅y denotes the Euclidean inner product of x and y. If C is generated by a polynomial g(x) satisfying g(x)∣xn−λ, and h(x) is given by h(x)=xn−λg(x), then h(x) is called the parity check polynomial of code C. It is a classical result that the dual code C⊥ is generated by h(x)∗, where h(x)∗=h(0)−1xdeg(h(x))h(x−1) is the reciprocal polynomial of h(x). The code C is called to be a self-orthogonal if C⊆C⊥ and a self-dual code if C=C⊥. For self-dual cyclic code, a well-known result states that there exist self-dual cyclic codes of length n over Fq if and only if n is even and the characteristic of Fq is p=2.
There are q−1 classes of constacyclic codes of length n over Fq. However, some of them are turned out to be equivalent in the sense that they have the same structure. To be explicit, two elements λ,μ∈F∗q are called n-equivalent in F∗qif there exists a∈F∗q such that anλ=μ.
Lemma 2.3. [5] For any λ,μ∈F∗q, the following four statements are equivalent:
(1) λ and μ are n-equivalent in F∗q.
(2) λ−1μ∈⟨ξn⟩.
(3) (λ−1μ)d=1, where d=q−1gcd(n,q−1).
(4) There exists an a∈F∗q such that
φa:Fq[X]/(Xn−μ)→Fq[X]/(Xn−λ);f(X)↦f(aX) |
is an Fq-algebra isomorphism. In particular, there are gcd(n,q−1) n-equivalence classes in F∗q.
We conclude this section with the introduction of q-cyclotomic coset which is important in the computation of constacyclic codes. Let n be a positive integer relatively prime to n. For 0≤s≤n−1, the q-cyclotomic coset of s modulo n is defined to be
Cs={s,sq,⋯,sqns−1}, |
where ns is the least positive integer such that sqns≡s(modn). It is obvious to see that ns is equal to the multiplicative order of q modulo ngcd(s,n). Notice that if sqa≡s′qb(modn) for some positive integers a,b, then
s≡sqa+(ns−a)≡s′qb+(ns−a)(modn). |
It follows that for 0≤s,s′≤n−1, Cs∩Cs′≠∅ if and only if Cs=Cs′. Therefore the q-cyclotomic cosets give a classification of the element in Zn.
If α is a primitive nth root of unit in some extension field of Fq, then the polynomial
Cs(x)=∏i∈Cs(x−αi) |
is exactly the minimal polynomial of αs over Fq, and
xn−1=∏sCs(x) |
gives the irreducible factorization of xn−1 over Fq, where s runs over all representations of distinct q-cyclotomic cosets modulo n. We call Cs(x) the polynomial associated to Cs.
Let Cs={s,sq,⋯,sqns−1} be any q-cyclotomic coset modulo n. The reciprocal coset of Cs is defined to be
C∗s={−s,−sq,⋯,−sqns−1}. |
We say that the coset Cs is self-reciprocal if Cs=C∗s. One can check that the polynomial C∗s(x) associated to the reciprocal coset C∗s is exactly the reciprocal polynomial of Cs(x).
The q-cyclotomic cosets modulo p1pt2 plays an important role in determining all the constacyclic codes of length p1pt2ps. In this section we consider a more general case that classifies all the q-cyclotomic cosets modulo pt11pt22, where p1 and p2 are two distinct odd prime numbers not dividing q, and t1,t2 are positive integers.
Let ℓ be a prime number not dividing q, and μ be a generator of the cyclic group Z∗ℓ. It is obvious that all the q-cyclotomic cosets modulo ℓ are given by C0={0} and
Ck={μk,μkq,⋯,μkqordℓ(q)−1}, 1≤k≤ℓ−1ordℓ(q). |
For different odd prime numbers p1 and p2, we claim that there exists an integer μ1 satisfying that:
(1) μ1 is a primitive root modulo pd1 for all d≥1; and
(2) μ1≡1(modp2).
We begin with a random primitive root η′1 modulo p1. If p21∤η′1p1−1−1, we let η1=η′1, otherwise we let η1=η′1+p1. It is trivial to see that η1 satisfies the condition gcd(ηp1−11−1p1,p1)=1. Let μ1=η1+(1−η1)pp2−11, then
μp1−11−1≡(η1+(1−η1)pp2−11)p1−1−1≡ηp1−11−1(modp21). |
It follows that
gcd(μp1−11−1p1,p1)=gcd(ηp1−11−1p1,p1)=1. |
Following Lemma 2.1, μ1 is a primitive root modulo pd1 for all d≥1 such that μ1≡1(modp2). By the symmetric argument, we can find an integer μ2 satisfying that
(1) μ2 is a primitive root modulo pd2 for all d≥1; and
(2) μ2≡1(modp1).
We fix such a pair of integers μ1 and μ2.
Theorem 3.1. Let p1 and p2 be two different odd prime numbers not dividing q, and t1 and t2 be positive integers. Then all the distinct q-cyclotomic cosets module pt11pt22 are given by
Cμk11μk22pr11pr22={μk11μk22pr11pr22,μk11μk22pr11pr22q,⋯,μk11μk22pr11pr22qcr1,r2} |
for 0≤r1≤t1, 0≤r2≤t2, 0≤k1≤g1,t1−r1−1 and 0≤k2≤g2,t2−r2⋅gcd(f1,t1−r1,f2,t2−r2)−1, where cr1,r2=ordpt1−r11pt2−r22(q)=lcm(f1,t1−r1,f2,t2−r2).
Proof. First we prove that the given q-cyclotomic cosets are all distinct. If Cμk11μk22pr11pr22=Cμk′11μk′22pr′11pr′22 for some 0≤r1,r′1≤t1, 0≤r2,r′2≤t2, 0≤k1,k′1≤g1,t1−r1−1 and 0≤k2,k′2≤g2,t2−r2⋅gcd(f1,t1−r1,f2,t2−r2)−1, then there exists a positive integer m such that
μk′11μk′22pr′11pr′22≡μk11μk22pr11pr22qm(modpt11pt22). | (3.1) |
Since μ1,μ2 and q are relatively prime to pt11pt22, clearly we have r1=r′1 and r2=r′2, and Eq (3.1) can be reduced to
μk′11μk′22≡μk11μk22qm(modpt1−r11pt2−r22). |
Remembering that μ1≡1(modp2) and μ2≡1(modp1), then by the Chinese remainder theorem, we have
μk1−k′11≡qm(modpt1−r11) | (3.2) |
μk2−k′22≡qm(modpt2−r22) | (3.3) |
Equation (3.2) implies that
μ(k1−k′1)f1,t1−r11≡qm⋅f1,t1−r1≡1(modpt1−r11), |
and therefore ϕ(pt1−r11)∣(k1−k′1)f1,t1−r1. Since 0≤k1,k′1≤g1,t1−r1−1, one must have k1=k′1. Notice that k1=k′1 indicates that qm≡1(modpt1−r11), then f1,t1−r1∣m, which together with Eq (3.3) leads to
μ(k′2−k2)⋅f2,t2−r2gcd(f1,t1−r1,f2,t2−r2)2≡qm⋅f2,t2−r2gcd(f1,t1−r1,f2,t2−r2)≡1(modpt2−r22). |
Thus ϕ(pt2−r22)∣(k′2−k2)⋅f2,t2−r2gcd(f1,t1−r1,f2,t2−r2). Since 0≤k2,k′2≤g2,t2−r2⋅gcd(f1,t1−r1,f2,t2−r2)−1, we have k2=k′2.
On the other hand, there are in total
∑0≤r1≤t1∑0≤r2≤t2ϕ(pt1−r11)f1,t1−r1⋅ϕ(pt2−r22)f2,t2−r2⋅gcd(f1,t1−r1,f2,t2−r2)⋅lcm(f1,t1−r1,f2,t2−r2)=∑0≤r1≤t1∑0≤r2≤t2ϕ(pt1−r11)ϕ(pt2−r22)=pt11pt22 | (3.4) |
elements in these q-cyclotomic cosets, therefore they are all the distinct q-cyclotomic cosets module pt11pt22.
In particular, when t1=1 and t2=t, the classification of the q-cyclotomic cosets modulo p1pt2 is given as follow.
Corollary 3.1. Let the notations be as above. Then all the distinct q-cyclotomic cosets modulo p1pt2 are
C0={0}; |
Cμk11μk22pr2={μk11μk22pr2,μk11μk22pr2q,⋯,μk11μk22pr2qordp1pt−r2(q)−1} |
for 0≤r≤t−1, 0≤k1≤g1−1 and 0≤k2≤g2,t−r⋅gcd(f1,f2,t−r);
Cμk1pt2={μk1pt2,μk1pt2q,⋯,μk1pt2qf1−1} |
for 0≤k≤g1−1; and
Cμk′2p1pr2={μk′2p1pr2,μk′2p1pr2q,⋯,μk′2p1pr2qf2,t−r−1} |
for 0≤r≤t−1 and 0≤k′≤g2,t−r−1.
Corollary 3.2. Let the notations be as aboved. Then the irreducible factorization of xp1pt2ps−1 over Fq is given by
xp1pt2ps−1=C0(x)pst−1∏r=0g1−1∏k1=0g2,t−rgcd(f1,f2,t−r)−1∏k2=0Cμk11μk22pr2(x)psg1−1∏k=0Cμk1pt2(x)pst−1∏r=0g2,t−r−1∏k′=0Cμk′2p1pr2(x)ps. |
In this section, we will determine the generator polynomials of all constacyclic codes of length p1pt2ps over Fq and their dual codes. For λ∈F∗q, we identify a λ-constacyclic code of length p1pt2ps with an ideal (g(x)) of the quotient ring Fq[x]/(xp1pt2ps−λ), where g(x) is a divisor of xp1pt2ps−λ. By Lemma 2.3, there are gcd(p1pt2,q−1) p1pt2ps-equivalence classes in F∗q, which corresponds to the cosets of ⟨ξp1pt2⟩ in F∗q=⟨ξ⟩.
Before doing the explicit computation, we present a general method to factorize xn−λ. Let q=pk for k>0, and n=pepe11⋯pemm be the prime factorization of n. Assume that pe11⋯pemm∣q−1, i.e., vpi(q−1)≥ei for i=1,⋯,m. In this case we have
F∗q=⟨ξ⟩=⟨ξpe11⋯pemm⟩∪⟨ξpe11⋯pemm⟩ξpe∪⋯∪⟨ξpe11⋯pemm⟩ξpe(pe11⋯pemm−1). |
For λ∈⟨ξpe11⋯pemm⟩ξj⋅pe, where 0≤j≤pe11⋯pemm−1, there exists an element a∈F∗q such that
anλ=ξj⋅pe. |
We first factorize xn−ξjpe, 0≤j≤pe11⋯pemm−1. Notice that j can be written as j=y⋅pv11⋯pvmm, where vi=min{ei,vpi(j)}. Then we have
xn−ξj⋅pe=(xpe11⋯pemm−ξy⋅pv11⋯pvmm)pe=ξj⋅pe((xpe1−v11⋯pem−vmmξy)pv11⋯pvmm−1)pe. |
Since pv11⋯pvmm∣q−1, δ=ξq−1pv11⋯pvmm is a primitive pv11⋯pvmm-th root of unit. Then
xn−ξj⋅pe=ξj⋅pe(xpe1−v11⋯pem−vmmξy−1)pe⋅(xpe1−v11⋯pem−vmmξy−δ)pe⋯(xpe1−v11⋯pem−vmmξy−δpv11⋯pvmm−1)pe=(xpe1−v11⋯pem−vmm−ξy)pe(xpe1−v11⋯pem−vmm−δξy)pe⋯(xpe1−v11⋯pem−vmm−δpv11⋯pvmm−1ξy)pe. |
For 0≤i≤pv11⋯pvmm−1, δiξy=ξy+i⋅q−1pv11⋯pvmm, and then we have
ord(δiξy)=q−1gcd(q−1,y+i⋅q−1pv11⋯pvmm), |
and
q−1ord(δiξy)=gcd(q−1,y+i⋅q−1pv11⋯pvmm). |
For each pi∣pe1−v11⋯pem−vmm, we have that ei>vi and vi=vpi(j), thus pi∤y. Since vpi(q−1)≥ei>vi, pi∣q−1pv11⋯pvmm, which indicates that pi∤y+i⋅q−1pv11⋯pvmm and pi∣q−1y+i⋅q−1pv11⋯pvmm. Moreover if 4∣pe1−v11⋯pem−vmm, then 4∣pe11⋯pemm∣q−1. Hence by Lemma 2.2 each xpe1−v11⋯pem−vmm−ξyδi is irreducible over Fq.
Notice that anλ=ξjpe, then the irreducible factorization of xn−λ follows immediately:
xn−λ=(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmξy)pe(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmδξy)pe⋅⋅⋯(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmδpv11⋯pvmm−1ξy)pe, |
We summerize the above discussions into the following theorem.
Theorem 4.1. Let p,p1,⋯,pm be distinct prime numbers. Let q=pk and n=pepe11⋯pemm, where k,e,e1,⋯,em are positive integers. Suppose that for 1≤i≤m, vpi(q−1)≥ei. Then for any λ∈F∗q, there exists an element a∈F∗q such that anλ=ξjpe, 0≤j≤pe11⋯pemm. Furthermore, writing j in the form j=y⋅pv11⋯pvmm, where vi=min{ei,vpi(j)}, then
xn−λ=(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmξy)pe(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmδξy)pe⋅⋅⋯(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmδpv11⋯pvmm−1ξy)pe, |
gives the irreducible factorization of xn−λ over Fq.
Now we turn to the case that pe11⋯pemm∤q−1. Sinve gcd(pe11⋯pemm,q)=1, thus there exists a least positive integer d such that pe11⋯pemm∣qd−1. By the lifting-the-exponent lemma, if d′ is the least positive integer such that p1⋯pm∣qd′−1, then d=d′pv11⋯pvmm, where vi=max{ei−vpi(qd′−1),0}.
Let λ be a nonzero element in Fq. To obtain the irreducible factorization of xn−λ over Fq, we first consider the factorization over Fqd. By Theorem 4.1, there exists a∈Fqd such that anλ=ζjpe, 0≤j≤pe11⋯pemm−1. Writing j as j=y⋅pv11⋯pvmm, where vi=min{ei,vpi(j)}, then
xn−λ=(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmζy)pe(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmδζy)pe⋅⋅⋯(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmδpv11⋯pvmm−1ζy)pe, |
gives the irreducible factorization of xn−λ over Fqd, where δ is a primitive pv11⋯pvmm-th root of unit. Hence each irreducible factor of xn−λ over Fq is of the form
(xpe1−v11⋯pem−vmm−a−pe1−v11⋯pem−vmmδiζy)pe(xpe1−v11⋯pem−vmm−a−qpe1−v11⋯pem−vmmδqiζqy)pe⋅⋅⋯(xpe1−v11⋯pem−vmm−a−qzi−1pe1−v11⋯pem−vmmδi⋅qzi−1ζy⋅qzi−1)pe, |
where zi is the least positive integer such that a−qzipe1−v11⋯pem−vmmδi⋅qziζy⋅qzi=a−pe1−v11⋯pem−vmmδiζy.
Now we determine the generator polynomials of all constacyclic codes of length p1pt2ps and their duals explicitly. We decompose the problem into three cases.
As gcd(q−1,p1pt2ps)=1, all constacyclic codes of length p1pt2ps are equivalent to a cyclic code. By the factorization of xp1pt2ps−1 given in Corollary 3.2, we have the following result. For any polynomial
F=a0+a1x+⋯+anxn, an≠0, |
we set ˆF=a−1nF to be the monic polynomial associated to F.
Proposition 4.1. Assume that gcd(q−1,p1pt2ps)=1. Then any nonzero element λ in Fq is p1pt2ps-equivalent to 1, that is, there is an element a∈F∗q such that ap1pt2psλ=1. Furthermore, the irreducible factorization of xp1pt2ps−λ over Fq is given by
xp1pt2ps−λ=ˆC0(ax)pst−1∏r=0g1−1∏k1=0g2,t−rgcd(f1,f2,t−r)−1∏k2=0ˆCμk11μk22pr2(ax)psg1−1∏k=0ˆCμk1pt2(ax)pst−1∏r=0g2,t−r−1∏k′=0ˆCμk′2p1pr2(ax)ps. |
Therefore all the constacyclic codes of length p1pt2ps are
C=(ˆC0(ax)ut−1∏r=0g1−1∏k1=0g2,t−rgcd(f1,f2,t−r)−1∏k2=0ˆCμk11μk22pr2(ax)vμk11μk22pr2g1−1∏k=0ˆCμk1pt2(ax)wμk1pt2t−1∏r=0g2,t−r−1∏k′=0ˆCμk′2p1pr2(ax)xμk′2p1pr2), |
where 0≤u,vμk11μk22pr2,wμk1pt2,xμk′2p1pr2≤ps, with duals
C⊥=(ˆC0(a−1x)ps−ut−1∏r=0g1−1∏k1=0g2,t−rgcd(f1,f2,t−r)−1∏k2=0ˆCμk11μk22pr2(a−1x)ps−vμk11μk22pr2g1−1∏k=0ˆCμk1pt2(a−1x)ps−wμk1pt2t−1∏r=0g2,t−r−1∏k′=0ˆCμk′2p1pr2(a−1x)ps−xμk′2p1pr2). |
For this case, since p_{1}p_{2}^{t}\vert q-1 , the following proposition follows straightly from Theorem 4.1.
Theorem 4.2. Assume that \gcd(q-1, p_{1}p_{2}^{t}p^{s}) = p_{1}p_{2}^{t} . Then for any \lambda\in \mathbb{F}_{q}^* , there exists an element a\in \mathbb{F}_{q}^* such that a^{p_{1}p_{2}^{t}p^{s}}\lambda = \xi^{j\cdot p^{s}} , 0\leq j\leq p_{1}p_{2}^{t}-1 . Writing j as j = y\cdot p_{1}^{v_{1}}p_{2}^{v_{2}} , where v_{1} = \min\{1, v_{p_{1}}(j)\} and v_{2} = \min\{t, v_{p_{2}}(j)\} , then
\begin{eqnarray*} x^{p_{1}p_{2}^{t}p^{s}}-\lambda& = &(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\xi^{y})^{p^{s}}(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta\xi^{y})^{p^{s}}\\ &&\cdots(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{p_{1}^{v_{1}}p_{2}^{v_{2}}-1}\xi^{y})^{p^{s}} \end{eqnarray*} |
gives the irreducible factorization of x^{p_{1}p_{2}^{t}p^{s}}-\lambda over \mathbb{F}_{q} . Therefore all the \lambda -constacyclic codes of length p_{1}p_{2}^{t}p^{s} and their dual codes are given by
\begin{eqnarray*} C& = &\left((x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\xi^{y})^{u_{1}}(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta\xi^{y})^{u_{2}}\right.\\ &&\left.\cdots(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{p_{1}^{v_{1}}p_{2}^{v_{2}}-1}\xi^{y})^{u _{p_{1}^{v_{1}}p_{2}^{v_{2}}}}\right), \end{eqnarray*} |
and
\begin{eqnarray*} C^{\bot}& = &\left((x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\xi^{-y})^{p^{s}-u_{1}}(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{-1}\xi^{-y})^{p^{s}-u_{2}}\right.\\ &&\left.\cdots(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{1-p_{1}^{v_{1}}p_{2}^{v_{2}}}\xi^{-y})^{p^{s}-u_{p_{1}^{v_{1}}p_{2}^{v_{2}}}}\right), \end{eqnarray*} |
where 0\leq u_{1}, u_{2}, \cdots, u_{n^{v_{1}}\ell^{v_{2}}}\leq p^{s} .
In this case, for any d\geq 1 we have f_{2, d} = p_{2}^{max\{0, d-r\}} , and f = \mathrm{lcm}(f_{1}, f_{2, t}) is the least positive integer such that q^{f}\equiv 1 \pmod {p_1p_{2}^{t}} . By the bais results of finite fields, there is a primitive element \zeta in \mathbb{F}_{q^{f}}^{*} such that \xi = \zeta^{\frac{q^{f}-1}{q-1}} = \zeta^{1+q+\cdots+q^{f-1}} . Then we have
\mathbb{F}_{q}^* = \langle\xi\rangle = \langle\xi^{p_{2}^{r}}\rangle\cup \langle\xi^{p_{2}^{r}}\rangle\xi^{p^{s}}\cup \cdots\cup \langle\xi^{p_{2}^{r}}\rangle\xi^{(p_{2}^{r}-1)p^{s}} |
and
\mathbb{F}_{q^{f}}^* = \langle\zeta\rangle = \langle\zeta^{p_{1}p_{2}^{t}}\rangle\cup \langle\zeta^{p_{1}p_{2}^{t}}\rangle\zeta^{p^{s}}\cup \cdots\cup \langle\zeta^{p_{1}p_{2}^{t}}\rangle\zeta^{(p_{1}p_{2}^{t} -1)p^{s}}. |
By the assumption that p_{1}p_{2}^{t}\mid q^{f}-1 and v_{p_{1}}(q-1) = 0 , v_{p_{2}}(q-1) = r , we have that p_{1}p_{2}^{t-r}\mid (1+q+\cdots+q^{f-1}) . Therefore \xi^{p_{2}^{r}} = \zeta^{p_{2}^{r}(1+q+\cdots+q^{f-1})}\in \langle\zeta^{p_{1}p_{2}^{t}}\rangle . Furthermore, for 0\leq j\leq p_{2}^{r}-1 , there exists some 0\leq j^{'}\leq p_{1}p_{2}^{t}-1 such that jp^{s}(1+q+\cdots+q^{f-1})\equiv j^{'}p^{s} \pmod {p_{1}p_{2}^{t}} , that is, \xi^{jp^{s}}\in \langle\zeta^{p_{1}p_{2}^{t}}\rangle\zeta^{j^{'}p^{s}} . Hence we have the following theorem.
Theorem 4.3. Assume that \gcd(q-1, p_{1}p_{2}^{t}p^{s}) = p_{2}^{r} , 0 < r \leq t . For any 0\leq j\leq p_{2}^{r}-1 , there exists an element a\in \mathbb{F}_{q^{f}}^{*} such that a^{p_{1}p_{2}^{t}p^{s}}\xi^{j\cdot p^{s}} = \zeta^{j'\cdot p^{s}} . Moreover, each irreducible factor of x^{p_{1}p_{2}^{t}}-\xi^{j} over \mathbb{F}_{q} is of the form
\begin{eqnarray*} &&(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{i}\zeta^{y'})(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
where j' = y'p_{1}^{v_{1}}p_{2}^{v_{2}} , v_{1} = \min\{1, v_{p_{1}}(j')\} , v_{2} = \min\{t, v_{p_{2}}(j')\} , and z_{i} is the least positive integer such that a^{-q^{z_{i}}p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{iq^{z_{i}}}\zeta^{y'q^{z_{i}}} = a^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{i}\zeta^{y'} .
For any 0\leq i, i^{'}\leq p_{1}^{v_{1}}p_{2}^{v_{2}}-1 , we define a relation \sim to be such that i\sim i^{'} if and only if a^{-q^{m}p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{iq^{m}}\zeta^{y'q^{m}} = a^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{i^{'}}\zeta^{y'} for some nonnegative integers m . It is obvious to see that \sim is an equivalence relation. Assume that S is a complete system of equivalence class representatives of \{0, 1, \cdots, p_{1}^{v_{1}}p_{•}^{v_{2}}-1\} relative to this relation \sim . For any i\in S we denote the irreducible polynomial
\begin{eqnarray*} &&(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{i}\zeta^{y'})(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
by M_{i}(x) . Then we have the following corollary.
Corollary 4.1. Assume that \gcd(q-1, p_{1}p_{2}^{t}p^{s}) = p_{2}^{r} . For any 0\leq j\leq p_{2}^{r}-1 , there exists an element a\in \mathbb{F}_{q^{f}}^{*} such that a^{p_{1}p_{2}^{t}p^{s}}\xi^{j\cdot p^{s}} = \zeta^{j'\cdot p^{s}} . Then
x^{p_{1}p_{2}^{t}p^{s}}-\xi^{j p^{s}} = \prod\limits_{i\in S}M_{i}(x)^{p^{s}} |
gives the irreducible factorization of x^{p_{1}p_{2}^{t}p^{s}}-\xi^{j p^{s}} over \mathbb{F}_q . Furthermore we have that
C = \left(\prod\limits_{i\in S}M_{i}(x)^{u_{i}}\right), |
and
C^{\bot} = \left(\prod\limits_{i\in S}M_{i}^{*}(x)^{p^{s}-u_{i}}\right), |
where 0\leq u_{i}\leq p^{s} , for i\in S .
The same argument as in the last section can be applied in this situation, only noticing that the least positive integer f such that q^{f}\equiv 1 \pmod {p_1p_2^t} is f = f_{2, t} = p_2^{\max\{0, t-r\}} . We find a primitive element \zeta in \mathbb{F}_{q^{f}}^{*} such that \xi = \zeta^{\frac{q^{f}-1}{q-1}} = \zeta^{1+q+\cdots+q^{f-1}} , then
\mathbb{F}_{q}^* = \langle\xi\rangle = \langle\xi^{p_{1}p_{2}^{r}}\rangle\cup \langle\xi^{p_{1}p_{2}^{r}}\rangle\xi^{p^{s}}\cup \cdots\cup \langle\xi^{p_{1}p_{2}^{r}}\rangle\xi^{(p_{2}^{r}-1)p^{s}} |
and
\mathbb{F}_{q^{f}}^* = \langle\zeta\rangle = \langle\zeta^{p_{1}p_{2}^{t}}\rangle\cup \langle\zeta^{p_{1}p_{2}^{t}}\rangle\zeta^{p^{s}}\cup \cdots\cup \langle\zeta^{p_{1}p_{2}^{t}}\rangle\zeta^{(p_{1}p_{2}^{t} -1)p^{s}}. |
By the assumption that p_{1}p_{2}^{t}\mid q^{f}-1 and v_{p_{2}}(q-1) = r , we have that p_{2}^{t-r}\mid (1+q+\cdots+q^{f-1}) , and \xi^{p_{1}p_{2}^{r}} = \zeta^{p_{1}p_{2}^{r}(1+q+\cdots+q^{f-1})}\in \langle\zeta^{p_{1}p_{2}^{t}}\rangle . Furthermore, for 0\leq j\leq p_{1}p_{2}^{r}-1 , there exists some 0\leq j^{'}\leq p_{1}p_{2}^{t}-1 such that jp^{s}(1+q+\cdots+q^{f-1})\equiv j^{'}p^{s} \pmod {p_1p_2^t} , that is, \xi^{jp^{s}}\in \langle\zeta^{p_{1}p_{2}^{t}}\rangle\zeta^{j^{'}p^{s}} . Hence we have the following theorem.
Theorem 4.4. Assume that \gcd(q-1, p_{1}p_{2}^{t}p^{s}) = p_{1}p_{2}^{r} for 0 < r < t , then for any 0\leq j\leq p_{2}^{r}-1 , there exists an element a\in \mathbb{F}_{q^{f}}^{*} such that a^{p_{1}p_{2}^{t}p^{s}}\xi^{j\cdot p^{s}} = \zeta^{j'\cdot p^{s}} . Moreover, each irreducible factor of x^{p_{1}p_{2}^{t}}-\xi^{j} over \mathbb{F}_{q} is of the form
\begin{eqnarray*} &&(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{i}\zeta^{y'})(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
where j' = y'p_{1}^{v_{1}}p_{2}^{v_{2}} , v_{1} = \min\{1, v_{p_{1}}(j')\} , v_{2} = \min\{t, v_{p_{2}}(j')\} , and z_{i} is the least positive integer such that a^{-q^{z_{i}}p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{iq^{z_{i}}}\zeta^{y'q^{z_{i}}} = a^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{i}\zeta^{y'} .
For any 0\leq i, i^{'}\leq p_{1}^{v_{1}}p_{2}^{v_{2}}-1 , we define a relation \sim to be such that i\sim i^{'} if and only if a^{-q^{m}p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{iq^{m}}\zeta^{y'q^{m}} = a^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{i^{'}}\zeta^{y'} for some nonnegative integers m . It is obvious to see that \sim is an equivalence relation. Assume that S is a complete system of equivalence class representatives of \{0, 1, \cdots, p_{1}^{v_{1}}p_2^{v_{2}}-1\} relative to this relation \sim . For any i\in S we denote the irreducible polynomial
\begin{eqnarray*} &&(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}\delta^{i}\zeta^{y'})(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}}-a^{-p_{1}^{1-v_{1}}p_{2}^{t-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
by M_{i}(x) . Then we have the following corollary.
Corollary 4.2. Assume that \gcd(q-1, p_{1}p_{2}^{t}p^{s}) = p_{1}p_{2}^{r} for 0 < r < t . For any 0\leq j\leq p_{2}^{r}-1 , there exists an element a\in \mathbb{F}_{q^{f}}^{*} such that a^{p_{1}p_{2}^{t}p^{s}}\xi^{j\cdot p^{s}} = \zeta^{j'\cdot p^{s}} . Then
x^{p_{1}p_{2}^{t}p^{s}}-\xi^{j p^{s}} = \prod\limits_{i\in Z}M_{i}(x)^{p^{s}} |
gives the irreducible factorization of x^{p_{1}p_{2}^{t}p^{s}}-\xi^{j p^{s}} over \mathbb{F}_q . Furthermore we have that
C = \left(\prod\limits_{i\in S}M_{i}(x)^{u_{i}}\right), |
and
C^{\bot} = \left(\prod\limits_{i\in S}M_{i}^{*}(x)^{p^{s}-u_{i}}\right), |
where 0\leq u_{i}\leq p^{s} , for i\in S .
Based on the results in the last section, we now give all the self-dual cyclic codes of length p_{1}p_{2}^{t}p^{s} over \mathbb{F}_q and their enumeration. It is a well-known conclusion that self-dual cyclic codes of length N over \mathbb{F}_q exists if and only if N is even and the characteristic of \mathbb{F}_q is 2 . Therefore we only consider the case of self-dual cyclic codes of length p_{1}p_{2}^{t}\cdot 2^{s} over \mathbb{F}_{2^{k}} .
Let x^{p_{1}p_{2}^{t}2^{s}}-1 = (x^{p_{1}p_{2}^{t}}-1)^{2^{s}} = f_{1}(x)^{2^{s}}\cdots f_{n}(x)^{2^{s}}h_{1}(x)^{2^{s}}\cdots h_{m}(x)^{2^{s}}h_{1}^{*}(x)^{2^{s}}\cdots h_{m}^{*}(x)^{2^{s}} be the irreducible factorization of x^{p_{1}p_{2}^{t}2^{s}}-1 over \mathbb{F}_q , where each f_{i}(x) is a monic irreducible self-reciprocal polynomial for 1\leq i\leq n , and h_{j}^{*}(x) is the reciprocal polynomial of h_{j}(x) for each 1\leq j\leq m . Now, given a cyclic code C = (g(x)) of length p_{1}p_{2}^{t}2^{s} , it can be written in the form
g(x) = f_{1}(x)^{\tau_{1}}\cdots f_{n}(x)^{\tau_{n}}h_{1}(x)^{\delta_{1}}\cdots h_{m}(x)^{\delta_{m}}h_{1}^{*}(x)^{\sigma_{1}}\cdots h_{m}^{*}(x)^{\sigma_{m}}, |
where 0\leq \tau_{i}, \delta_{j}, \sigma_{j}\leq 2^{s} for any 1\leq i\leq n and 1\leq j\leq m . Then the reciprocal polynomial h^{*}(x) of the parity check polynomial h(x) of C is
h^{*}(x) = f_{1}(x)^{2^{s}-\tau_{1}}\cdots f_{n}(x)^{2^{s}-\tau_{n}}h_{1}(x)^{2^{s}-\sigma_{1}}\cdots h_{m}(x)^{2^{s}-\sigma_{m}}h_{1}^{*}(x)^{2^{s}-\delta_{1}}\cdots h_{m}^{*}(x)^{2^{s}-\delta_{m}}. |
Therefore it is obvious to see that the following theorem holds.
Theorem 5.1. With the above notations, we have that C is self-dual if and only if 2\tau_{i} = 2^{s} for 1\leq i\leq n , and \delta_{j}+\sigma_{j} = 2^{s} for 1\leq j\leq m .
Recall the irreducible factorization of x^{p_{1}p_{2}^{t}p^{s}}-1 given in Corollary 3.2. Now we determine for each irreducible factor its reciprocal polynomial.
Lemma 5.1. Let the notations be defined as Corollary 3.1. Then one of the following holds.
(1) If both f_{1} and f_{2} are odd, then we have that
C_{0}^{*} = C_{0}, \ C_{\mu_{1}^{k}p_{2}^{t}}^{*} = C_{-\mu_{1}^{k}p_{2}^{t}}, \ C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}^{*} = C_{-\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}, \ C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}^{*} = C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}. |
(2) If f_{1} is odd and f_{2} is even, then we have that
C_{0}^{*} = C_{0}, \ C_{\mu_{1}^{k}p_{2}^{t}}^{*} = C_{-\mu_{1}^{k}p_{2}^{t}}, \ C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}^{*} = C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}, \ C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}^{*} = C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}. |
(3) If f_{1} is even and f_{2} is odd, then we have that
C_{0}^{*} = C_{0}, \ C_{\mu_{1}^{k}p_{2}^{t}}^{*} = C_{\mu_{1}^{k}p_{2}^{t}}, \ C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}^{*} = C_{-\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}, \ C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}^{*} = C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}. |
(4) If both f_{1} and f_{2} are even, then we have when v_{2}(f_{1})\neq v_{2}(f_{2}) ,
C_{0}^{*} = C_{0}, \ C_{\mu_{1}^{k}p_{2}^{t}}^{*} = C_{\mu_{1}^{k}p_{2}^{t}}, \ C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}^{*} = C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}, \ C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}^{*} = C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}, |
when v_{2}(f_{1}) = v_{2}(f_{2}) ,
C_{0}^{*} = C_{0}, \ C_{\mu_{1}^{k}p_{2}^{t}}^{*} = C_{\mu_{1}^{k}p_{2}^{t}}, \ C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}^{*} = C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}, \ C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}^{*} = C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}. |
Proof. First it is trivial that the reciprocal of C_{0} is always itself. For C_{\mu_{1}^{k}p_{2}^{t}} , notice that C_{\mu_{1}^{k}p_{2}^{t}}^{*} = C_{\mu_{1}^{k}p_{2}^{t}} if and only if the congruence equation -\mu_{1}^{k}p_{2}^{t}\equiv -\mu_{1}^{k}p_{2}^{t}q^{x} \pmod {p_1p_{2}^{t}} is solvable. Since the equation is equivalent to -1\equiv q^{x} \pmod {p_2^{r}} , then the condition holds if and only if f_{1} = \mathrm{ord}_{p_{1}}(q) is even. In the similar way we can check that C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}^{*} = C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}} if and only if f_{2, t-r} = f_{2}p_{2}^{max\{0, t-r\}} is even. Notice that by assumption p_{2} is odd, therefore the condition holds if and only if f_{2} is even. For C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}} , consider the congruence equation -\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}\equiv \mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}q^{x} \pmod {p_1p_{2}^{t}} . It is equivalent to that -1\equiv q^{x} \pmod {p_1} and -1\equiv q^{x} \pmod {p_2^{t-r}} holds simultaneously. This requires not only both f_{1} and f_{2} are even, but also \gcd(f_{1}, f_{2, t-r})\mid\dfrac{f_{1}-f_{2, t-r}}{2} . And it is trivial to check that the last condition holds if and only if v_{2}(f_{1}) = v_{2}(f_{2, t-r}) = v_{2}(f_{2}) .
Based on the above lemma, we now determine all the self-dual cyclic codes of length p_{1}p_{2}^{t} and their enumeration.
Theorem 5.2.
(1) If both f_{1} and f_{2} are odd, then there exist (2^{s}+1)^{\frac{p_{1}p_{2}^{t}-1}{2}} self-dual cyclic codes of length p_{1}p_{2}^{t} over \mathbb{F}_{2^{k}} , which are given by
\begin{eqnarray*} &\left((x-1)^{2^{s-1}}\prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k_{1} = 0}^{\frac{g_{1}}{2}-1}\prod\nolimits_{k_{2} = 0}^{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})-1} C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{v_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}} C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{2^{s}-v_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}} \right.\\ &\left.\cdot\prod\nolimits_{k = 0}^{\frac{g_{1}}{2}-1} C_{\mu_{1}^{k}p_{2}^{t}}(x)^{w_{\mu_{1}^{k}p_{2}^{t}}} C_{-\mu_{1}^{k}p_{2}^{t}}(x)^{2^{s}-w_{\mu_{1}^{k}p_{2}^{t}}} \prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k^{\prime} = 0}^{\frac{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})}{2}-1} C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}(x)^{x_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}} C_{-\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}(x)^{2^{s}-x_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}}\right). \end{eqnarray*} |
(2) If f_{1} is odd and f_{2} is even, then there exist (2^{s}+1)^{\frac{p_{1}(p_{2}^{t}-1)}{2}} self-dual cyclic codes of length p_{1}p_{2}^{t} over \mathbb{F}_{2^{k}} , which are given by
\begin{eqnarray*} &\left((x-1)^{2^{s-1}}\prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k_{1} = 0}^{\frac{g_{1}}{2}-1}\prod\nolimits_{k_{2} = 0}^{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})-1} C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{v_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}} C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{2^{s}-v_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}} \right.\\ &\left.\cdot\prod\nolimits_{k = 0}^{\frac{g_{1}}{2}-1} C_{\mu_{1}^{k}p_{2}^{t}}(x)^{w_{\mu_{1}^{k}p_{2}^{t}}} C_{-\mu_{1}^{k}p_{2}^{t}}(x)^{2^{s}-w_{\mu_{1}^{k}p_{2}^{t}}} \prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k^{\prime} = 0}^{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})-1} C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}(x)^{2^{s-1}} \right). \end{eqnarray*} |
(3) If f_{1} is even and f_{2} is odd, then there exist (2^{s}+1)^{\frac{p_{2}^{t}(p_{1}-1)}{2}} self-dual cyclic codes of length p_{1}p_{2}^{t} over \mathbb{F}_{2^{m}} , which are given by
\begin{eqnarray*} &\left((x-1)^{2^{s-1}}\prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k_{1} = 0}^{g_{1}-1}\prod\nolimits_{k_{2} = 0}^{\frac{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})}{2}-1} C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{v_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}} C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{2^{s}-v_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}} \right.\\ &\left.\cdot\prod\nolimits_{k = 0}^{g_{1}-1} C_{\mu_{1}^{k}p_{2}^{t}}(x)^{2^{s-1}} \prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k^{\prime} = 0}^{\frac{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})}{2}-1} C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}(x)^{x_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}} C_{-\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}(x)^{2^{s}-x_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}}\right). \end{eqnarray*} |
(4) If both f_{1} and f_{2} are even, then we have when v_{2}(f_{1})\neq v_{2}(f_{2}) , there exist (2^{s}+1)^{\frac{(p_{1}-1)(p_{2}^{t}-1)}{2}} self-dual cyclic codes of length p_{1}p_{2}^{t} over \mathbb{F}_{2^{m}} , which are given by
\begin{eqnarray*} &\left((x-1)^{2^{s-1}}\prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k_{1} = 0}^{g_{1}-1}\prod\nolimits_{k_{2} = 0}^{\frac{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})}{2}-1} C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{v_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}} C_{-\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{2^{s}-v_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}} \right.\\ &\left.\cdot\prod\nolimits_{k = 0}^{g_{1}-1} C_{\mu_{1}^{k}p_{2}^{t}}(x)^{2^{s-1}} \prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k^{\prime} = 0}^{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})-1} C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}(x)^{2^{s-1}} \right). \end{eqnarray*} |
When v_{2}(f_{1}) = v_{2}(f_{2}) , there exist only one self-dual cyclic codes of length p_{1}p_{2}^{t} over \mathbb{F}_{2^{m}} , which is given by
\begin{eqnarray*} &\left((x-1)^{2^{s-1}}\prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k_{1} = 0}^{g_{1}-1}\prod\nolimits_{k_{2} = 0}^{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})-1} C_{\mu_{1}^{k_{1}}\mu_{2}^{k_{2}}p_{2}^{r}}(x)^{2^{s-1}} \right.\\ &\left.\cdot\prod\nolimits_{k = 0}^{g_{1}-1} C_{\mu_{1}^{k}p_{2}^{t}}(x)^{2^{s-1}} \prod\nolimits_{r = 0}^{t-1}\prod\nolimits_{k^{\prime} = 0}^{g_{2, t-r}\gcd(f_{1}, f_{2, t-r})-1} C_{\mu_{2}^{k^{\prime}}p_{1}p_{2}^{r}}(x)^{2^{s-1}} \right). \end{eqnarray*} |
In this section, we illustrate the above process with the example of constacyclic codes of length 5\ell p^{s} , where \ell is a prime number different from 5 and p . We determine all the constacyclic codes of length 5\ell p^{s} and their dual codes over \mathbb{F}_{q} , and then all the self-dual codes of length 5\ell p^{s} are also given.
First we determine all the q -cyclotomic cosets modulo 5\ell . Let f = \mathrm{ord}_{\ell}(q) , and e = \dfrac{\ell-1}{f} . Then we have:
(1) \mathrm{ord}_{5\ell}(q) = f , when q\equiv 1 \pmod 5 .
(2) \mathrm{ord}_{5\ell}(q) = f , when q\equiv 4 \pmod 5 with f even.
(3) \mathrm{ord}_{5\ell}(q) = 2f , when q\equiv 4 \pmod 5 with f odd.
(4) \mathrm{ord}_{5\ell}(q) = f , when q\equiv 2 or q\equiv 3 \pmod 5 with 4\mid f .
(5) \mathrm{ord}_{5\ell}(q) = 2f , when q\equiv 2 or q\equiv 3 \pmod 5 with 2\mid f but 4\nmid f .
(6) \mathrm{ord}_{5\ell}(q) = 4f , when q\equiv 2 or q\equiv 3 \pmod 5 with f odd.
As the discussion given in the Section 3, we can find a primitive root \mu modulo \ell^{t} for all t\geq 1 such that \mu\equiv 1 \pmod 5 . The following lemma give more explicit formula for the q -cyclotomic cosets modulo 5\ell .
Lemma 6.1.
(1) If q\equiv 1 \pmod 5 , then we have that all the distinct q -cyclotomic cosets modulo 5\ell are given by C_{0} = \{0\} , C_{\ell} = \{\ell\} , C_{2\ell} = \{2\ell\} , C_{-\ell} = \{-\ell\} , C_{-2\ell} = \{-2\ell\} , and C_{a\mu^{k}} = \{a\mu^{k}, a\mu^{k}q, \cdots, a\mu^{k}q^{f-1}\} for a\in R = \{1, 2, -1, -2, 5\} and 0\leq k\leq e-1 .
(2) If q\equiv 4 \pmod 5 and f is even, we have that all the distinct q -cyclotomic cosets modulo 5\ell are given by C_{0} = \{0\} , C_{\ell} = \{\ell, \ell q\} , C_{2\ell} = \{2\ell, 2\ell q\} , C_{\mu^{k'}} = \{\mu^{k'}, \mu^{k'}q, \cdots, \mu^{k'}q^{f-1}\} , C_{2\mu^{k'}} = \{2\mu^{k'}, 2\mu^{k'}q, \cdots, 2\mu^{k'}q^{f-1}\} for 0\leq k'\leq 2e-1 , and C_{5\mu^{k}} = \{5\mu^{k}, 5\mu^{k}q, \cdots, 5\mu^{k}q^{f-1}\} for 0\leq k\leq e-1 .
(3) If q\equiv 4 \pmod 5 and f is odd, we have that all the distinct q -cyclotomic cosets modulo 5\ell are given by C_{0} = \{0\} , C_{\ell} = \{\ell, \ell q\} , C_{2\ell} = \{2\ell, 2\ell q\} , C_{\mu^{k}} = \{\mu^{k}, \mu^{k}q, \cdots, \mu^{k}q^{2f-1}\} , C_{2\mu^{k}} = \{2\mu^{k}, 2\mu^{k}q, \cdots, 2\mu^{k}q^{2f-1}\} , and C_{5\mu^{k}} = \{5\mu^{k}, 5\mu^{k}q, \cdots, 5\mu^{k}q^{f-1}\} for 0\leq k\leq e-1 .
(4) If q\equiv 2 or 3 \pmod 5 and 4\mid f , we have that all the distinct q -cyclotomic cosets modulo 5\ell are given by C_{0} = \{0\} , C_{\ell} = \{\ell, \ell q, \ell q^{2}, \ell q^{3}\} , C_{\mu^{k'}} = \{\mu_{k'}, \mu^{k'}q, \cdots, \mu^{k'}q^{f-1}\} for 0\leq k'\leq 4e-1 , and C_{5\mu^{k}} = \{5\mu^{k}, 5\mu^{k}q, \cdots, 5\mu^{k}q^{f-1}\} for 0\leq k\leq e-1 .
(5) If q\equiv 2 or 3 \pmod 5 and 2\mid f but 4\nmid f , we have that all the distinct q -cyclotomic cosets modulo 5\ell are given by C_{0} = \{0\} , C_{\ell} = \{\ell, \ell q, \ell q^{2}, lq^{3}\} , C_{\mu^{k'}} = \{\mu_{k'}, \mu^{k'}q, \cdots, \mu^{k'}q^{2f-1}\} for 0\leq k'\leq 2e-1 , and C_{5\mu^{k}} = \{5\mu^{k}, 5\mu^{k}q, \cdots, 5\mu^{k}q^{f-1}\} for 0\leq k\leq e-1 .
(6) If q\equiv 2 or 3 \pmod 5 and f is odd, we have that all the distinct q -cyclotomic cosets modulo 5\ell are given by C_{0} = \{0\} , C_{\ell} = \{\ell, \ell q, \ell q^{2}, \ell q^{3}\} , C_{\mu^{k}} = \{\mu_{k}, \mu^{k}q, \cdots, \mu^{k}q^{4f-1}\} , and C_{5\mu^{k}} = \{5\mu^{k}, 5\mu^{k}q, \cdots, 5\mu^{k}q^{f-1}\} for 0\leq k\leq e-1 .
Proof. The methods to prove the above 6 situations are similar, and we will give the proof of the second situation as a instance. First since \mu is a fixed primitive root modulo l such that \mu\equiv 1 \pmod 5 , it is trivial to verify that C_{0} , C_{\ell} , C_{2\ell} , C_{\mu^{k'}} , C_{2\mu^{k'}} for 0\leq k'\leq 2e-1 and C_{5\mu^{k}} for 0\leq k\leq e-1 are q -cyclotomic cosets modulo 5\ell . And then we claim that all these cosets are all distinct. If we have that a_{1}\mu^{k_{1}}\equiv a_{2}\mu^{k_{2}}q^{j} , where a_{1} , a_{2} , k_{1} , k_{2} and j satisfy the definitions in (2). Since
\gcd(a_{1}, 5\ell ) = \gcd(a_{1}\mu^{k_{1}}, 5\ell ) = \gcd(a_{2}\mu^{k_{2}}q^{j}, 5\ell ) = \gcd(a_{2}, 5\ell ), |
we have that either a_{1} = a_{2} or a_{1}\neq a_{2} and both a_{1} and a_{2} are not equal to 5 . We divide the proof into 2 subcases.
Subcase 1. If a_{1} = a_{2} , we have that \mu^{k_{1}-k_{2}}\equiv q^{j}\pmod {\ell} and \mu^{(k_{1}-k_{2})f}\equiv 1\pmod {\ell} , therefore \phi(\ell)\mid (k_{1}-k_{2})f and \dfrac{\phi(\ell)}{f}\mid (k_{1}-k_{2}) , which indicates that k_{1} = k_{2} .
Subcase 2. If a_{1}\neq a_{2} and none of them is equal to 5 , we have that a_{1}a_{2}^{-1}\equiv \mu^{k_{2}-k_{1}}q^{j}\pmod {5\ell} , but notice that a_{1}a_{2}^{-1}\equiv \pm2 \pmod 5 and \mu^{k_{2}-k_{1}}q^{j}\equiv \pm1 \pmod 5 , which is a contradiction. Hence the given cosets are all distinct, and we only need to prove they are all the q -cyclotomic cosets to complete the proof.
Notice that
\vert C_{0}\vert+\vert C_{\ell}\vert+\vert C_{2\ell}\vert+\sum\limits_{k' = 0}^{2e-1}\vert C_{\mu^{k'}}\vert+\sum\limits_{k' = 0}^{2e-1}\vert C_{2\mu^{k'}}\vert+\sum\limits_{k = 0}^{e-1}\vert C_{5\mu^{k}}\vert = 5+2ef+2ef+ef = 5(ef+1) = 5(\phi(\ell)+1) = 5\ell. |
Therefore the conclusion holds.
Theorem 6.1. The irreducible factorization of x^{5\ell }-1 over \mathbb{F}_q is given as follows.
(1) If q\equiv 1 \pmod 5 , then
x^{5\ell }-1 = C_{0}(x)C_{\ell}(x)C_{2\ell}(x)C_{3\ell}(x)C_{4\ell}(x)\prod\limits_{a\in R}\prod\limits_{k = 0}^{e-1}C_{a\mu^{k}}(x), |
where R = {1, 2, 3, 4, 5} .
(2) If q\equiv 4 \pmod 5 and f is even, then
x^{5\ell }-1 = C_{0}(x)C_{\ell}(x)C_{2\ell}(x)\prod\limits_{k' = 0}^{2e-1}C_{\mu^{k'}}(x)C_{2\mu^{k'}}(x)\prod\limits_{k = 0}^{e-1}C_{5\mu^{k}}(x), |
(3) If q\equiv 4 \pmod 5 and f is odd, then
x^{5\ell }-1 = C_{0}(x)C_{\ell}(x)C_{2\ell}(x)\prod\limits_{k = 0}^{e-1}C_{\mu^{k}}(x)C_{2\mu^{k}}(x)C_{5\mu^{k}}(x), |
(4) If q\equiv 2 or 3 \pmod 5 and 4\mid f , then
x^{5\ell }-1 = C_{0}(x)C_{\ell}(x)\prod\limits_{k' = 0}^{4e-1}C_{\mu^{k'}}(x)\prod\limits_{k = 0}^{e-1}C_{5\mu^{k}}(x), |
(5) If q\equiv 2 or 3 \pmod 5 and 2\mid f but 4\nmid f , then
x^{5\ell }-1 = C_{0}(x)C_{\ell}(x)\prod\limits_{k' = 0}^{2e-1}C_{\mu^{k'}}(x)\prod\limits_{k = 0}^{e-1}C_{5\mu^{k}}(x), |
(6) If q\equiv 2 or 3 \pmod 5 and f is odd, then
x^{5\ell }-1 = C_{0}(x)C_{\ell}(x)\prod\limits_{k = 0}^{e-1}C_{\mu^{k}}(x)C_{5\mu^{k}}(x), |
With the irreducible factorization of x^{5\ell }-1 , we can straightly follow the process given in Section 4 to calculate all the constacyclic codes of length 5\ell p^{s} over \mathbb{F}_{q} . We list the result as follow.
Theorem 6.2. Assume that \gcd(q-1, 5\ell p^{s}) = 1 , then \lambda -constacyclic codes C of length 5\ell p^{s} over \mathbb{F}_q are equivalent to the cyclic codes, i.e., for any \lambda\in \mathbb{F}_{q}^* , there exists a unique element a\in \mathbb{F}_{q}^* such that a^{5\ell p^{s}}\lambda = 1 . Furthermore, the irreducible factorization of x^{5\ell p^{s}}-\lambda over \mathbb{F}_q is given by
(1) If q\equiv 4 \pmod 5 and f is even, then
x^{5\ell p^{s}}-\lambda = \widehat{C}_{0}(ax)^{p^{s}}\widehat{C}_{\ell}(ax)^{p^{s}}\widehat{C}_{2\ell}(ax)^{p^{s}}\prod\limits_{k' = 0}^{2e-1}\widehat{C}_{\mu^{k'}} (ax)^{p^{s}}\widehat{C}_{2\mu^{k'}}(ax)^{p^{s}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{5\mu^{k}}(ax)^{p^{s}}, |
Therefore we have that
C = \left(\widehat{C}_{0}(ax)^{\varepsilon_{1}}\widehat{C}_{\ell}(ax)^{\varepsilon_{2}}\widehat{C}_{2\ell}(ax)^{\varepsilon_{3}}\prod\limits_{k' = 0}^{2e-1}\widehat{C}_{\mu^{k'}} (ax)^{\tau_{k^{'}}}\widehat{C}_{2\mu^{k'}}(ax)^{\nu_{k^{'}}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{5\mu^{k}}(ax)^{\rho_{k}}\right), |
and
\begin{eqnarray*} C^{\bot}& = &\left(\widehat{C}_{0}(a^{-1}x)^{p^{s}-\varepsilon_{1}}\widehat{C}_{-\ell}(a^{-1}x)^{p^{s}-\varepsilon_{2}}\widehat{C}_{-2\ell}(a^{-1}x)^{p^{s}-\varepsilon_{3}}\right.\\ &&\times\left.\prod\limits_{k' = 0}^{2e-1}\widehat{C}_{-\mu^{k'}} (a^{-1}x)^{p^{s}-\tau_{k^{'}}}\widehat{C}_{-2\mu^{k'}}(a^{-1}x)^{p^{s}-\nu_{k^{'}}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{-5\mu^{k}}(a^{-1}x)^{p^{s}-\rho_{k}}\right), \end{eqnarray*} |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \varepsilon_{3}, \tau_{k^{'}}, \nu_{k^{'}}, \rho_{k}\leq p^{s} , for any k^{'} = 0, 1, \cdots, 2e-1 , and k = 0, 1, \cdots, e-1 .
(2) If q\equiv 4 \pmod 5 and f is odd, then
x^{5\ell p^{s}}-\lambda = \widehat{C}_{0}(ax)^{p^{s}}\widehat{C}_{\ell}(ax)^{p^{s}}\widehat{C}_{2\ell}(ax)^{p^{s}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{\mu^{k}}(ax)^{p^{s}}\widehat{C}_{2\mu^{k}} (ax)^{p^{s}}\widehat{C}_{5\mu^{k}}(ax)^{p^{s}}. |
Therefore we have that
C = \left(\widehat{C}_{0}(ax)^{\varepsilon_{1}}\widehat{C}_{\ell}(ax)^{\varepsilon_{2}}\widehat{C}_{2\ell}(ax)^{\varepsilon_{3}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{\mu^{k}}(ax)^{\tau_{k}}\widehat{C}_{2\mu^{k}} (ax)^{\nu_{k}}\widehat{C}_{5\mu^{k}}(ax)^{\rho_{k}}\right), |
and
\begin{eqnarray*} C^{\bot}& = &\left(\widehat{C}_{0}(a^{-1}x)^{p^{s}-\varepsilon_{1}}\widehat{C}_{-\ell}(a^{-1}x)^{p^{s}-\varepsilon_{2}}\widehat{C}_{-2\ell}(a^{-1}x)^{p^{s}-\varepsilon_{3}}\right.\\ &&\times\left.\prod\limits_{k = 0}^{e-1}\widehat{C}_{-\mu^{k}}(a^{-1}x)^{p^{s}-\tau_{k}}\widehat{C}_{-2\mu^{k}} (a^{-1}x)^{p^{s}-\nu_{k}}\widehat{C}_{-5\mu^{k}}(a^{-1}x)^{p^{s}-\rho_{k}}\right), \end{eqnarray*} |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \varepsilon_{3}, \tau_{k}, \nu_{k}, \rho_{k}\leq p^{s} , for k = 0, 1, \cdots, e-1 .
(3) If q\equiv 2 or 3 \pmod 5 and 4\mid f , then
x^{5\ell p^{s}}-\lambda = \widehat{C}_{0}(ax)^{p^{s}}\widehat{C}_{\ell}(ax)^{p^{s}}\prod\limits_{k' = 0}^{4e-1}\widehat{C}_{\mu^{k'}}(ax)^{p^{s}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{5\mu^{k}}(ax)^{p^{s}}. |
Therefore we have that
C = \left(\widehat{C}_{0}(ax)^{\varepsilon_{1}}\widehat{C}_{\ell}(ax)^{\varepsilon_{2}}\prod\limits_{k' = 0}^{4e-1}\widehat{C}_{\mu^{k'}}(ax)^{\tau_{k^{'}}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{5\mu^{k}}(ax)^{\nu_{k}}\right), |
and
C^{\bot} = \left(\widehat{C}_{0}(a^{-1}x)^{p^{s}-\varepsilon_{1}}\widehat{C}_{-\ell}(a^{-1}x)^{p^{s}-\varepsilon_{2}}\prod\limits_{k' = 0}^{4e-1}\widehat{C}_{-\mu^{k'}}(a^{-1}x)^{p^{s}-\tau_{k^{'}}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{-5\mu^{k}}(a^{-1}x)^{p^{s}-\nu_{k}}\right), |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \tau_{k^{'}}, \nu_{k}\leq p^{s} , for k^{'} = 0, 1, \cdots, 4e-1 , and k = 0, 1, \cdots, e-1 .
(4) If q\equiv 2 or 3 \pmod 5 and 2\mid f but 4\nmid f , then
x^{5\ell p^{s}}-\lambda = \widehat{C}_{0}(ax)^{p^{s}}\widehat{C}_{\ell}(ax)^{p^{s}}\prod\limits_{k' = 0}^{2e-1}\widehat{C}_{\mu^{k'}}(ax)^{p^{s}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{5\mu^{k}}(ax)^{p^{s}}. |
Therefore we have that
C = \left(\widehat{C}_{0}(ax)^{\varepsilon_{1}}\widehat{C}_{\ell}(ax)^{\varepsilon_{2}}\prod\limits_{k' = 0}^{2e-1}\widehat{C}_{\mu^{k'}}(ax)^{\tau_{k^{'}}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{5\mu^{k}}(ax)^{\nu_{k}}\right), |
and
C^{\bot} = \left(\widehat{C}_{0}(a^{-1}x)^{p^{s}-\varepsilon_{1}}\widehat{C}_{-\ell}(a^{-1}x)^{p^{s}-\varepsilon_{2}}\prod\limits_{k' = 0}^{2e-1}\widehat{C}_{-\mu^{k'}}(a^{-1}x)^{p^{s}-\tau_{k^{'}}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{-5\mu^{k}}(a^{-1}x)^{p^{s}-\nu_{k}}\right), |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \tau_{k^{'}}, \nu_{k}\leq p^{s} , for k^{'} = 0, 1, \cdots, 2e-1 , and k = 0, 1, \cdots, e-1 .
(5) If q\equiv 2 or 3 \pmod 5 and f is odd, then
x^{5\ell p^{s}}-\lambda = \widehat{C}_{0}(ax)^{p^{s}}\widehat{C}_{\ell}(ax)^{p^{s}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{\mu^{k}}(ax)^{p^{s}}\widehat{C}_{5\mu^{k}}(ax)^{p^{s}}. |
Therefore we have that
C = \left(\widehat{C}_{0}(ax)^{\varepsilon_{1}}\widehat{C}_{\ell}(ax)^{\varepsilon_{2}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{\mu^{k}}(ax)^{\tau_{k}}\widehat{C}_{5\mu^{k}}(ax)^{\nu_{k}}\right), |
and
C^{\bot} = \left(\widehat{C}_{0}(a^{-1}x)^{p^{s}-\varepsilon_{1}}\widehat{C}_{-\ell}(a^{-1}x)^{p^{s}-\varepsilon_{2}}\prod\limits_{k = 0}^{e-1}\widehat{C}_{-\mu^{k}}(a^{-1}x)^{p^{s}-\tau_{k}}\widehat{C}_{-5\mu^{k}}(a^{-1}x)^{p^{s}-\nu_{k}}\right), |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \tau_{k}, \nu_{k}\leq p^{s} , for k = 0, 1, \cdots, e-1 .
Theorem 6.3. Assume that \gcd(q-1, 5\ell p^{s}) = 5\ell , then \mathbb{F}_{q}^* = \langle\xi\rangle = \langle\xi^{5\ell }\rangle\cup \langle\xi^{5\ell }\rangle\xi^{p^{s}}\cup \cdots\cup \langle\xi^{5\ell }\rangle\xi^{p^{s}(5\ell -1)} . For any \lambda\in \mathbb{F}_{q}^* , there exists an element a\in \mathbb{F}_{q}^* such that a^{5\ell p^{s}}\lambda = \xi^{j\cdot p^{s}} , where 0\leq j\leq 5\ell -1 . Then j can be written as j = y\cdot 5^{v_{1}}\ell^{v_{2}} , where v_{1} = \min\{1, v_{5}(j)\} and v_{2} = \min\{1, v_{\ell}(j)\} . And
\begin{eqnarray*} x^{n}-\lambda& = &(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\xi^{y})^{p^{s}}(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta\xi^{y})^{p^{s}}\\ &&\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{5^{v_{1}}\ell^{v_{2}}-1}\xi^{y})^{p^{s}} \end{eqnarray*} |
gives the irreducible factorization of x^{5\ell p^{s}}-\lambda over \mathbb{F}_{q} . Moreover, all the \lambda -constacyclic codes of length 5lp^{s} and their dual codes are given by
\begin{eqnarray*} C& = &\left((x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\xi^{y})^{\varepsilon_{1}}(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta\xi^{y})^{\varepsilon_{2}}\right.\\ &&\left.\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{5^{v_{1}}\ell^{v_{2}}-1}\xi^{y})^{\varepsilon_{5^{v_{1}}\ell^{v_{2}}}}\right), \end{eqnarray*} |
and
\begin{eqnarray*} C^{\bot}& = &\left((x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}}\xi^{-y})^{p^{s}-\varepsilon_{1}}(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{-1}\xi^{-y})^{p^{s}-\varepsilon_{2}}\right.\\ &&\left.\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{1-5^{v_{1}}\ell^{v_{2}}}\xi^{-y})^{p^{s}-\varepsilon_{5^{v_{1}}\ell^{v_{2}}}}\right), \end{eqnarray*} |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \cdots, \varepsilon_{5^{v_{1}}\ell^{v_{2}}}\leq p^{s} .
Theorem 6.4. Assume that \gcd(q-1, 5\ell p^{s}) = 5 , then for any 0\leq j\leq 4 , there exists an element a\in \mathbb{F}_{q^{f}}* such that a^{5\ell p^{s}}\xi^{j\cdot p^{s}} = \zeta^{j'\cdot p^{s}} . Moreover, each irreducible factor of x^{5\ell}-\xi^{j} over \mathbb{F}_{q} is of the form
\begin{eqnarray*} &&(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i}\zeta^{y'})(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
where j' = y'5^{v_{1}}\ell^{v_{2}} , v_{1} = \min\{1, v_{5}(j')\} , v_{2} = \min\{1, v_{\ell}(j')\} , and z_{i} is the least positive integer such that a^{-q^{z_{i}}5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{iq^{z_{i}}}\zeta^{y'q^{z_{i}}} = a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i}\zeta^{y'} .
For any 0\leq i, i^{'}\leq 5^{v_{1}}\ell^{v_{2}}-1 , we define a relation \sim to be such that i\sim i^{'} if and only if a^{-q^{m}5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{iq^{m}}\zeta^{y'q^{m}} = a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i^{'}}\zeta^{y'} for some nonnegative integers m . It is obvious to see that \sim is an equivalence relation. Assume that S is a complete system of equivalence class representatives of \{0, 1, \cdots, 5^{v_{1}}\ell^{v_{2}}-1\} relative to this relation \sim . For any i\in S we denote the irreducible polynomial
\begin{eqnarray*} &&(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i}\zeta^{y'})(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
by M_{i}(x) , and denote
\begin{eqnarray*} &&(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{-i}\zeta^{-y'})(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}\cdot q}\delta^{-iq}\zeta^{-y'q})\\ &&\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}\cdot q^{z_{i}-1}}\delta^{-iq^{z_{i}-1}}\zeta^{-y'q^{z_{i}-1}}), \end{eqnarray*} |
by M_{i}^{'}(x) . Then we have the following corollary.
Corollary 6.1. Assume that \gcd(q-1, 5\ell p^{s}) = 5 . For any 0\leq j\leq 4 , there exists an element a\in \mathbb{F}_{q^{f}}* such that a^{5\ell p^{s}}\xi^{j\cdot p^{s}} = \zeta^{j'\cdot p^{s}} . Then
x^{5\ell p^{s}}-\xi^{j p^{s}} = \prod\limits_{i\in S}M_{i}(x)^{p^{s}} |
gives the irreducible factorization of x^{5\ell p^{s}}-\xi^{j p^{s}} over \mathbb{F}_q . Furthermore we have that
C = \left(\prod\limits_{i\in X}M_{i}(x)^{\varepsilon_{i}}\right), |
and
C^{\bot} = \left(\prod\limits_{i\in X}M_{i}^{'}(x)^{p^{s}-\varepsilon_{i}}\right), |
where 0\leq \varepsilon_{i}\leq p^{s} , for i\in X .
Theorem 6.5. Assume that \gcd(q-1, 5\ell p^{s}) = \ell , then
(1) If q\equiv 4 \pmod 5 , for any 0\leq j\leq \ell-1 , the following equations
j'\equiv 2j\pmod{\ell}\;\mathit{\text{ and }}\;j'\equiv 0 \pmod 5 |
have a unique solution j' up to modulo 5\ell . Moreover, each irreducible facotor of x^{5\ell}-\xi^{j} over \mathbb{F}_{q} is of the form
\begin{eqnarray*} &&(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i}\zeta^{y'})(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
where j' = y'5^{v_{1}}\ell^{v_{2}} , v_{1} = \min\{1, v_{5}(j')\} , v_{2} = \min\{1, v_{\ell}(j')\} , and z_{i} is the least positive integer such that a^{-q^{z_{i}}5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{iq^{z_{i}}}\zeta^{y'q^{z_{i}}} = a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i}\zeta^{y'} .
(2) If q\equiv 2, 3 \pmod 5 , for any 0\leq j\leq \ell-1 , the following equations
j'\equiv 4j \pmod{\ell} |
j'\equiv 0 \pmod 5 |
have a unique solution j' up to modulo 5\ell . Moreover, each irreducible facotor of x^{5\ell}-\xi^{j} over \mathbb{F}_{q} is of the form
\begin{eqnarray*} &&(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i}\zeta^{y'})(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
where j' = y'5^{v_{1}}\ell^{v_{2}} , v_{1} = min{1, v_{5}(j')} , v_{2} = min{1, v_{\ell}(j')} , and z_{i} is the least positive integer such that a^{-q^{z_{i}}5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{iq^{z_{i}}}\zeta^{y'q^{z_{i}}} = a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i}\zeta^{y'} .
For any 0\leq i, i^{'}\leq 5^{v_{1}}\ell^{v_{2}}-1 , we define a relation \sim to be such that i\sim i^{'} if and only if a^{-q^{m}5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{iq^{m}}\zeta^{y'q^{m}} = a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i^{'}}\zeta^{y'} for some nonnegative integer m . It is obvious to see that \sim is an equivalence relation. Assume that S is a complete system of equivalence class representatives of \{0, 1, \cdots, 5^{v_{1}}\ell^{v_{2}}-1\} relative to this relation \sim . For any i\in S we denote the irreducible polynomial
\begin{eqnarray*} &&(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{i}\zeta^{y'})(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q}\delta^{iq}\zeta^{y'q})\\ &&\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{-5^{1-v_{1}}\ell^{1-v_{2}}\cdot q^{z_{i}-1}}\delta^{iq^{z_{i}-1}}\zeta^{y'q^{z_{i}-1}}), \end{eqnarray*} |
by M_{i}(x) , and denote
\begin{eqnarray*} &&(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}}\delta^{-i}\zeta^{-y'})(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}\cdot q}\delta^{-iq}\zeta^{-y'q})\\ &&\cdots(x^{5^{1-v_{1}}\ell^{1-v_{2}}}-a^{5^{1-v_{1}}\ell^{1-v_{2}}\cdot q^{z_{i}-1}}\delta^{-iq^{z_{i}-1}}\zeta^{-y'q^{z_{i}-1}}), \end{eqnarray*} |
by M_{i}^{'}(x) .
Corollary 6.2. Assume that \gcd(q-1, 5\ell p^{s}) = \ell , then
(1) If q\equiv 4 \pmod 5 , and j, j^{'} is defined as in the first case of Theorem 5.1, then
x^{5\ell p^{s}}-\xi^{j p^{s}} = \prod\limits_{i\in X}M_{i}(x)^{p^{s}} |
gives the irreducible factorization of x^{5\ell p^{s}}-\xi^{j p^{s}} over \mathbb{F}_q . Furthermore we have that
C = \left(\prod\limits_{i\in X}M_{i}(x)^{\varepsilon_{i}}\right), |
and
C^{\bot} = \left(\prod\limits_{i\in X}M_{i}^{'}(x)^{p^{s}-\varepsilon_{i}}\right), |
where 0\leq \varepsilon_{i}\leq p^{s} , for i\in X .
(2) If q\equiv 2, 3 \pmod 5 , and j, j^{'} is defined as in the second case of Theorem 5.1, then
x^{5\ell p^{s}}-\xi^{j p^{s}} = \prod\limits_{i\in X}M_{i}(x)^{p^{s}} |
gives the irreducible factorization of x^{5\ell p^{s}}-\xi^{j p^{s}} over \mathbb{F}_q . Furthermore we have that
C = \left(\prod\limits_{i\in X}M_{i}(x)^{\varepsilon_{i}}\right), |
and
C^{\bot} = \left(\prod\limits_{i\in X}M_{i}^{'}(x)^{p^{s}-\varepsilon_{i}}\right), |
where 0\leq \varepsilon_{i}\leq p^{s} , for i\in X .
Finally we give all the self-dual constacylic codes of length 5\ell p^{s} as the end of this section. Since self-dual cyclic codes of length N over \mathbb{F}_q exists if and only if N is even and the characteristic of \mathbb{F}_q is p = 2 , as in the general case, we only consider the case of self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} .
Lemma 6.2. Assume that q\equiv 1 \pmod 5 . For the q -cyclotomic cosets, one of the following holds.
(1) If f = \mathrm{ord}_{\ell}(q) is even, we have that
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{-\ell}, \ C_{2\ell}^{*} = C_{-2\ell}, \ C_{\mu^{k}}^{*} = C_{-\mu^{k}}, \ C_{2\mu^{k}}^{*} = C_{-2\mu^{k}}, \ C_{5\mu^{k}}^{*} = C_{5\mu^{k}}, |
where 0\leq k\leq e-1 .
(2) If f = \mathrm{ord}_{\ell}(q) is odd, we have that
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{-\ell}, \ C_{2\ell}^{*} = C_{-2\ell}, \ C_{\mu^{k}}^{*} = C_{-\mu^{k}}, \ C_{2\mu^{k}}^{*} = C_{-2\mu^{k}}, \ C_{5\mu^{k^{'}}}^{*} = C_{-5\mu^{k^{'}}}, |
where \{C_{5\mu^{k}}\} = \{C_{5\mu^{k^{'}}}\}\bigcup \{C_{-5\mu^{k^{'}}}\} , and 0\leq k\leq e-1 , 0\leq k^{'}\leq \dfrac{e}{2}-1 .
Proof.
(1) By the definition of reciprocal coset, it is clear that C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{-\ell}, \ C_{2\ell}^{*} = C_{-2\ell}, \ C_{\mu^{k}}^{*} = C_{-\mu^{k}}, \ C_{2\mu^{k}}^{*} = C_{-2\mu^{k}} , thus it remains to prove C_{5\mu^{k}}^{*} = C_{5\mu^{k}} . Let t = \frac{f}{2} . Since f = \mathrm{ord}_{\ell}(q) , it is trivial to see that q^{t}\equiv -1 \pmod \ell , and therefore we have that -5\mu^{k}\equiv 5\mu^{k}q^{t} \pmod {5\ell} . It follows immediately that C_{5\mu^{k}}^{*} = C_{5\mu^{k}} , for 0\leq k\leq e-1 .
(2) As in the first case, the conclusions that C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{-\ell}, \ C_{2\ell}^{*} = C_{-2\ell}, \ C_{\mu^{k}}^{*} = C_{-\mu^{k}}, \ C_{2\mu^{k}}^{*} = C_{-2\mu^{k}} are clear, and now we prove that C_{5\mu^{k^{'}}}^{*} = C_{-5\mu^{k^{'}}} . To see this, we claim that for any 0\leq k_{1}^{'}, k_{2}^{'}\leq \frac{e}{2}-1 , C_{5\mu^{k_{1}^{'}}}\neq C_{-5\mu^{k_{2}^{'}}} , and \{C_{5\mu^{k}}\} = \{C_{5\mu^{k^{'}}}\}\bigcup \{C_{-5\mu^{k^{'}}}\} . Assume that C_{5\mu^{k_{1}^{'}}} = C_{-5\mu^{k_{2}^{'}}} for some 0\leq k_{1}^{'}, k_{2}^{'}\leq \frac{e}{2}-1 , then we have that 5\mu^{k_{1}^{'}}\equiv -5\mu^{k_{2}^{'}}q^{j} \pmod {5\ell} for some 0\leq j\leq f-1 , which indicates that -\mu^{k_{1}^{'}-k_{2}^{'}}\equiv q^{j} \pmod \ell . Notice that f is odd, therefore we have that -\mu^{f(k_{1}^{'}-k_{2}^{'})}\equiv q^{jf} \equiv 1 \pmod \ell and \mu^{f(k_{1}^{'}-k_{2}^{'})}\equiv -1 \pmod \ell . It follows that \mu^{2f(k_{1}^{'}-k_{2}^{'})}\equiv 1 \pmod \ell , hence \phi(\ell)\mid 2f(k_{1}^{'}-k_{2}^{'}) and \frac{e}{2}\mid k_{1}^{'}-k_{2}^{'} . Since by the condition we have 0\leq k_{1}^{'}, k_{2}^{'}\leq \frac{e}{2}-1 , we deduce that k_{1}^{'} = k_{2}^{'} . Then the equation 5\mu^{k_{1}^{'}}\equiv -5\mu^{k_{2}^{'}}q^{j} \pmod {5\ell} can be reduced to -1\equiv q^{j} \pmod \ell . However, notice that \mathrm{ord}_{\ell}(q) = f is odd, such a positive integer j cannot exist, which is a contradiction. According to this, we have that for any 0\leq k_{1}^{'}, k_{2}^{'}\leq \frac{e}{2}-1 , C_{5\mu^{k_{1}^{'}}}\neq C_{-5\mu^{k_{2}^{'}}} . By comparing the number of elements, it is trivial to verify that \{C_{5\mu^{k}}\} = \{C_{5\mu^{k^{'}}}\}\bigcup \{C_{-5\mu^{k^{'}}}\} holds. Then by the definition of reciprocal coset, one immediately get that C_{5\mu^{k^{'}}}^{*} = C_{-5\mu^{k^{'}}} .
With the same method we can prove the results for the rest of cases. The proofs will be omitted.
Lemma 6.3. Assume that q\equiv 4 \pmod 5 . For the q -cyclotomic cosets, one of the following holds.
(1) If f = 2t is even, then
(i) when t is even, we have that
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{\ell}, \ C_{2\ell}^{*} = C_{2\ell}, \ C_{\mu^{k}}^{*} = C_{-\mu^{k}}, \ C_{2\mu^{k}}^{*} = C_{-2\mu^{k}}, \ C_{5\mu^{k}}^{*} = C_{5\mu^{k}}, |
where \{C_{\mu^{k^{'}}}\} = \{C_{\mu^{k}}\}\bigcup \{C_{-\mu^{k}}\} , \{C_{2\mu^{k^{'}}}\} = \{C_{2\mu^{k}}\}\bigcup \{C_{-2\mu^{k}}\} , for 0\leq k\leq e-1 , 0\leq k^{'}\leq 2e-1 .
(ii) If t is odd, we have that
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{\ell}, \ C_{2\ell}^{*} = C_{2\ell}, \ C_{\mu^{k^{'}}}^{*} = C_{\mu^{k^{'}}}, \ C_{2\mu^{k^{'}}}^{*} = C_{2\mu^{k^{'}}}, \ C_{5\mu^{k}}^{*} = C_{5\mu^{k}}, |
where 0\leq k\leq e-1 , 0\leq k^{'}\leq 2e-1 .
(2) when f is odd, then
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{\ell}, \ C_{2\ell}^{*} = C_{2\ell}, \ C_{\mu^{k^{'}}}^{*} = C_{-\mu^{k^{'}}}, \ C_{2\mu^{k^{'}}}^{*} = C_{-2\mu^{k^{'}}}, \ C_{5\mu^{k^{'}}}^{*} = C_{-5\mu^{k^{'}}}, |
where \{C_{\mu^{k}}\} = \{C_{\mu^{k^{'}}}\}\bigcup \{C_{-\mu^{k^{'}}}\} , \{C_{2\mu^{k}}\} = \{C_{2\mu^{k^{'}}}\}\bigcup \{C_{-2\mu^{k^{'}}}\} , \{C_{5\mu^{k}}\} = \{C_{5\mu^{k^{'}}}\}\bigcup \{C_{-5\mu^{k^{'}}}\} , for 0\leq k\leq e-1, 0\leq k^{'}\leq \frac{e}{2}-1 .
Lemma 6.4. Assume that q\equiv 2 or 3 \pmod 5 . For the q -cyclotomic cosets, one of the following holds.
(1) If 4\mid f . Let f = 4t , then
(i) when t is even, we have that
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{\ell}, \ C_{\mu^{k^{''}}}^{*} = C_{-\mu^{k^{''}}}, \ C_{5\mu^{k}}^{*} = C_{5\mu^{k}}, |
where \{C_{\mu^{k^{'}}}\} = \{C_{\mu^{k^{''}}}\}\bigcup \{C_{-\mu^{k^{''}}}\} , for 0\leq k\leq e-1 , 0\leq k^{''}\leq 2e-1 and 0\leq k^{'}\leq 4e-1 .
(ii) If t is odd, we have that
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{\ell}, \ C_{\mu^{k^{'}}}^{*} = C_{\mu^{k^{'}}}, \ C_{5\mu^{k}}^{*} = C_{5\mu^{k}}, |
where 0\leq k\leq e-1 , 0\leq k^{'}\leq 4e-1 .
(2) If 2\mid f but 4\nmid f , then
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{\ell}, \ C_{\mu^{k}}^{*} = C_{-\mu^{k}}, \ C_{5\mu^{k}}^{*} = B_{5\mu^{k}}, |
where \{C_{\mu^{k^{'}}}\} = \{C_{\mu^{k}}\}\bigcup \{C_{-\mu^{k}}\} , for 0\leq k\leq e-1 , 0\leq k^{'}\leq 2e-1 .
(3) If f is odd, then
C_{0}^{*} = C_{0}, \ C_{\ell}^{*} = C_{\ell}, \ C_{\mu^{k^{'}}}^{*} = C_{-\mu^{k^{'}}}, \ C_{5\mu^{k^{'}}}^{*} = C_{-5\mu^{k^{'}}}, |
where \{C_{\mu^{k}}\} = \{C_{\mu^{k^{'}}}\}\bigcup \{C_{-\mu^{k^{'}}}\} , \{C_{5\mu^{k}}\} = \{C_{5\mu^{k^{'}}}\}\bigcup \{C_{-5\mu^{k^{'}}}\} , for 0\leq k^{'}\leq \frac{e}{2}-1 , 0\leq k\leq e-1 .
From the above lemmas, we give all the self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} and their enumeration in the following theorems.
Theorem 6.6. Let q\equiv 1 \pmod 5 , then one of the following holds.
(1) If f = \mathrm{ord}_{\ell}(q) is even, there exist (2^{s}+1)^{2+2e} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which are given by
\begin{eqnarray*} &&\left((x-1)^{2^{s-1}}C_{\ell}(x)^{\varepsilon_{1}}C_{-\ell}(x)^{2^{s}-\varepsilon_{1}}C_{2\ell}(x)^{\varepsilon_{2}}C_{-2\ell}(x)^{2^{s}-\varepsilon_{2}}\right.\\ &&\times\left.\prod\limits_{k = 0}^{e-1}C_{\mu^{k}}(x)^{\tau_{k}}C_{-\mu^{k}}(x)^{2^{s}-\tau_{k}}C_{2\mu^{k}}(x)^{\rho_{k}}C_{-2\mu^{k}}(x)^{2^{s}-\rho_{k}}C_{5\mu^{k}}(x)^{2^{s-1}}\right), \end{eqnarray*} |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \tau_{k}, \rho_{k}\leq 2^{s} , for any 0\leq k\leq e-1 .
(2) If f = \mathrm{ord}_{\ell}(q) is odd, there exist (2^{s}+1)^{2+\dfrac{5e}{2}} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which are given by
\begin{eqnarray*} &&\left((x-1)^{2^{s-1}}C_{\ell}(x)^{\varepsilon_{1}}C_{-\ell}(x)^{2^{s}-\varepsilon_{1}}C_{2\ell}(x)^{\varepsilon_{2}}C_{-2\ell}(x)^{2^{s}-\varepsilon_{2}}\right.\\ &&\cdot\left.\prod\limits_{k = 0}^{e-1}C_{\mu^{k}}(x)^{\tau_{k}}C_{-\mu^{k}}(x)^{2^{s}-\tau_{k}}C_{2\mu^{k}}(x)^{\rho_{k}}C_{-2\mu^{k}}(x)^{2^{s}-\rho_{k}}\prod\limits_{k^{'} = 0}^{\frac{e}{2}-1}C_{5\mu^{k^{'}}}(x)^{\iota_{k^{'}}}C_{-5\mu^{k^{'}}}(x)^{2^{s}-\iota_{k^{'}}}\right), \end{eqnarray*} |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \tau_{k}, \rho_{k}, \iota_{k^{'}}\leq 2^{s} , for any 0\leq k\leq e-1 and any 0\leq k^{'}\leq \dfrac{e}{2}-1 .
Proof.
(1) By Lemma 6.2, any self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} has the form of
\begin{eqnarray*} &&\left((x-1)^{2^{s-1}}C_{\ell}(x)^{\varepsilon_{1}}C_{-\ell}(x)^{2^{s}-\varepsilon_{1}}C_{2\ell}(x)^{\varepsilon_{2}}C_{-2\ell}(x)^{2^{s}-\varepsilon_{2}}\right.\\ &&\times\left.\prod\limits_{k = 0}^{e-1}C_{\mu^{k}}(x)^{\tau_{k}}C_{-\mu^{k}}(x)^{2^{s}-\tau_{k}}C_{2\mu^{k}}(x)^{\rho_{k}}C_{-2\mu^{k}}(x)^{2^{s}-\rho_{k}}C_{5\mu^{k}}(x)^{2^{s-1}}\right), \end{eqnarray*} |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \tau_{k}, \rho_{k}\leq 2^{s} , for any 0\leq k\leq e-1 . Since each of \varepsilon_{1}, \varepsilon_{2} and \tau_{k}, \rho_{k} , 0\leq k\leq e-1 , has 2^{s}+1 possible values, we have in total (2^{s}+1)^{2+2e} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} .
(2) By Lemma 6.2, any self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} has the form of
\begin{eqnarray*} &&\left((x-1)^{2^{s-1}}C_{\ell}(x)^{\varepsilon_{1}}C_{-\ell}(x)^{2^{s}-\varepsilon_{1}}C_{2\ell}(x)^{\varepsilon_{2}}C_{-2\ell}(x)^{2^{s}-\varepsilon_{2}}\right.\\ &&\cdot\left.\prod\limits_{k = 0}^{e-1}C_{\mu g^{k}}(x)^{\tau_{k}}C_{-\mu^{k}}(x)^{2^{s}- \tau_{k}}C_{2\mu^{k}}(x)^{\rho_{k}}C_{-2\mu^{k}}(x)^{2^{s}-\rho_{k}}\prod\limits_{k^{'} = 0}^{\frac{e}{2}-1}C_{5\mu^{k^{'}}}(x)^{\iota_{k^{'}}}C_{-5\mu^{k^{'}}}(x)^{2^{s}-\iota_{k^{'}}}\right), \end{eqnarray*} |
where 0\leq \varepsilon_{1}, \varepsilon_{2}, \tau_{k}, \rho_{k}, \iota_{k^{'}}\leq 2^{s} , for any 0\leq k\leq e-1 and any 0\leq k^{'}\leq \dfrac{e}{2}-1 . Each of \varepsilon_{1}, \varepsilon_{2} , \tau_{k}, \rho_{k}, 0\leq k\leq e-1 , and \iota_{k^{'}} , 0\leq k^{'}\leq \dfrac{e}{2}-1 , has 2^{s}+1 possible values, we have in total (2^{s}+1)^{2+\dfrac{5e}{2}} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} .
The proofs of theorems for the rest of cases are similar, and we will give them without proofs.
Theorem 6.7. Let q\equiv 4 \pmod 5 , then one of the following holds.
(1) If f = 2t is even, then
(i) when t is even, there exist (2^{s}+1)^{2e} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which are given by
\left((x-1)^{2^{s-1}}C_{\ell}(x)^{2^{s-1}}C_{2\ell}(x)^{2^{s-1}}\prod\limits_{k = 0}^{e-1}C_{\mu^{k}}(x)^{\tau_{k}}C_{-\mu^{k}}(x)^{2^{s}-\tau_{k}}C_{2g^{k}}(x)^{\rho_{k}}C_{-2g^{k}}(x)^{2^{s}-\rho_{k}}C_{5g^{k}}(x)^{2^{s-1}}\right), |
where 0\leq \tau_{k}, \rho_{k}\leq 2^{s} , for any 0\leq k\leq e-1 .
(ii) when t is odd, there exists only one self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which is given by
\left((x-1)^{2^{s-1}}C_{\ell}(x)^{2^{s-1}}C_{2\ell}(x)^{2^{s-1}}\prod\limits_{k^{'} = 0}^{2e-1}C_{\mu^{k^{'}}}(x)^{2^{s-1}}C_{2\mu^{k^{'}}}(x)^{2^{s-1}}\prod\limits_{k = 0}^{e-1}C_{5\mu^{k}}(x)^{2^{s-1}}\right). |
(2) If f is odd, thenthere exist (2^{s}+1)^{3e/2} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which are given by
\begin{eqnarray*} &&\left((x-1)^{2^{s-1}}C_{\ell}(x)^{2^{s-1}}C_{2\ell}(x)^{2^{s-1}}\right.\\ &&\times\left.\prod\limits_{k^{'} = 0}^{e/2-1}C_{\mu^{k^{'}}}(x)^{\tau_{k^{'}}}C_{-\mu^{k^{'}}}(x)^{2^{s}-\tau_{k^{'}}}C_{2\mu^{k^{'}}}(x)^{\rho_{k^{'}}}C_{-2\mu^{k^{'}}}(x)^{2^{s}-\rho_{k^{'}}}C_{5\mu^{k^{'}}}(x)^{\iota_{k^{'}}}C_{-5\mu^{k^{'}}}(x)^{2^{s}-\iota_{k^{'}}}\right). \end{eqnarray*} |
Theorem 6.8. Let q\equiv 2 or 3 \pmod 5 , then one of the following holds.
(1) If 4\mid f . Let f = 4t , then
(i) when t is even, there exist (2^{s}+1)^{2e} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which are given by
\left((x-1)^{2^{s-1}}C_{\ell}(x)^{2^{s-1}}\prod\limits_{k^{''} = 0}^{2e-1}C_{\mu^{k}}(x)^{\tau_{k^{''}}}C_{-\mu^{k^{''}}}(x)^{2^{s}-\tau_{k^{''}}}\prod\limits_{k = 0}^{e-1}C_{5\mu^{k}}(x)^{2^{s-1}}\right), |
where 0\leq \tau_{k^{''}}\leq 2^{s} , for any 0\leq k^{''}\leq 2e-1 .
(ii) when t is odd, there exists only one self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which is given by
\left((x-1)^{2^{s-1}}C_{\ell}(x)^{2^{s-1}}\prod\limits_{k^{'} = 0}^{4e-1}C_{\mu^{k^{'}}}(x)^{2^{s-1}}\prod\limits_{k = 0}^{e-1}C_{5\mu^{k}}(x)^{2^{s-1}}\right), |
(2) If 2\mid f but 4\nmid f , then there exist (2^{s}+1)^{e} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which are given by
\left((x-1)^{2^{s-1}}C_{\ell}(x)^{2^{s-1}}\prod\limits_{k = 0}^{e-1}C_{\mu^{k}}(x)^{\tau_{k}}C_{-\mu^{k}}(x)^{2^{s}-\tau_{k}}C_{5\mu^{k}}(x)^{2^{s-1}}\right), |
where 0\leq \tau_{k}\leq 2^{s} , for any 0\leq k\leq e-1 .
(3) If f is odd, then there exist (2^{s}+1)^{e} self-dual cyclic codes of length 5\cdot 2^{s}\ell over \mathbb{F}_{2^{m}} , which are given by
\left((x-1)^{2^{s-1}}C_{\ell}(x)^{2^{s-1}}\prod\limits_{k^{'} = 0}^{\dfrac{e}{2}-1}C_{\mu^{k^{'}}}(x)^{\tau_{k^{'}}}C_{-\mu^{k^{'}}}(x)^{2^{s}-\tau_{k^{'}}}C_{5\mu^{k^{'}}}(x)^{\iota_{k^{'}}}C_{-5\mu^{k^{'}}}(x)^{2^{s}-\iota_{k^{'}}}\right), |
where 0\leq \tau_{k^{'}}, \iota_{k^{'}}\leq 2^{s} , for any 0\leq k^{'}\leq \dfrac{e}{2}-1 .
The first author was supported by the Yuyou Team Support Program of North China University of Technology (No. 107051360019XN137/007) and Yujie Talent Project of North China University of Technology(No. 107051360022XN735).
The authors declare no conflict of interest.
[1] | L. Lakatos, L. Szeidl, M. Telek, Introduction to Queueing Systems with Telecommunication Applications, 2 Eds., Springer Publishing Company, Incorporated, 2019. |
[2] | X. Y. Jiang, K. Hong, Z. W. Fu, Skew cyclic displacements and decompositions of inverse matrix for an innovative structure matrix, J. Nonlinear Sci. Appl., 10 (2017), 4058–4070. http://dx.doi.org/10.22436/jnsa.010.08.02 |
[3] | A. Böttcher, B. Silbermann, Analysis of Toeplitz Operators, 2 Eds., Springer-Verlag Berlin Heidelberg, 2019. |
[4] |
Y. Q. Bai, T. Z. Huang, X. M. Gu, Circulant preconditioned iterations for fractional diffusion equations based on Hermitian and skew-Hermitian splittings, Appl. Math. Lett., 48 (2015), 14–22. http://dx.doi.org/10.1016/j.aml.2015.03.010 doi: 10.1016/j.aml.2015.03.010
![]() |
[5] |
M. K. Ng, J. Pan, Weighted Toeplitz regularized least squares computation for image restoration, SIAM J. Sci. Comput., 36 (2014), B94–B121. http://dx.doi.org/10.1137/120888776 doi: 10.1137/120888776
![]() |
[6] |
Z. Z. Bai, G. Q. Li, L. Z. Lu, Combinative preconditioners of modified incomplete Cholesky factorization and Sherman-Morrison-Woodbury update for self-adjoint elliptic Dirichlet-periodic boundary value problems, J. Comput. Math., 22 (2004), 833–856. http://dx.doi.org/doi:10.1016/j.cam.2004.02.011 doi: 10.1016/j.cam.2004.02.011
![]() |
[7] |
Z. Z. Bai, Z. R. Ren, Block-triangular preconditioning methods for linear third-order ordinary differential equations based on reduced-order sinc discretizations, J. Industr. Appl. Math., 30 (2013), 511–527. http://dx.doi.org/10.1007/s13160-013-0112-6 doi: 10.1007/s13160-013-0112-6
![]() |
[8] |
Z. Z. Bai, R. H. Chan, Z. R. Ren, On sinc discretization and banded preconditioning for linear third-order ordinary differential equations, Numer. Linear Algebra Appl., 18 (2011), 471–497. https://doi.org/10.1002/nla.738 doi: 10.1002/nla.738
![]() |
[9] |
Z. Z. Bai, R. H. Chan, Z. R. Ren, On order-reducible sinc discretizations and block-diagonal preconditioning methods for linear third-order ordinary differential equations, Numer. Linear Algebra Appl., 21 (2014), 108–135. http://dx.doi.org/10.1002/nla.1868 doi: 10.1002/nla.1868
![]() |
[10] | M. Shi, F. özbudak, L. Xu, P. Solé, LCD codes from tridiagonal Toeplitz matrices, Finite Fields Appl., 75 (2021), 101892. https://linkinghub.elsevier.com/retrieve/pii/S1071579721000861 |
[11] | M. Shi, L. Xu, P. Solé, On isodual double Toeplitz codes, 2021. https://arXiv.org/pdf/2102.09233v1 |
[12] |
C. F. Cao, S. Huang, The commutants of analytic Toeplitz operators for several complex variables, Sci. China Math., 53 (2010), 1877–1884. http://dx.doi.org/10.1007/s11425-010-4023-6 doi: 10.1007/s11425-010-4023-6
![]() |
[13] |
X. F. Wang, G. F. Cao, J. Xia, Toeplitz operators on Fock-Sobolev spaces with positive measure symbols, Sci. China Math., 57 (2014), 1443–1462. http://dx.doi.org/10.1007/s11425-014-4813-3 doi: 10.1007/s11425-014-4813-3
![]() |
[14] |
J. Y. Yang, Y. F. Lu, Commuting dual Toeplitz operators on the harmonic Bergman space, Sci. China Math., 58 (2015), 1461–1472. http://dx.doi.org/10.1007/s11425-014-4940-x doi: 10.1007/s11425-014-4940-x
![]() |
[15] |
X. F. Zhao, D. C. Zheng, The spectrum of Bergman Toeplitz operators with some harmonic symbols, Sci. China Math., 59 (2016), 731–740. https://doi.org/10.1007/s11425-015-5083-4 doi: 10.1007/s11425-015-5083-4
![]() |
[16] |
G. X. Ji, Analytic Toeplitz algebras and the Hilbert transform associated with a subdiagonal algebra, Sci. China Math., 57 (2014), 579–588. https://doi.org/10.1007/s11425-013-4684-z doi: 10.1007/s11425-013-4684-z
![]() |
[17] |
M. K. Ng, K. Rost, Y. W. Wen, On inversion of Toeplitz matrices, Linear Algebra Appl., 348 (2002), 145–151. https://doi.org/10.1016/S0024-3795(01)00592-4 doi: 10.1016/S0024-3795(01)00592-4
![]() |
[18] |
G. Labahn, T. Shalom, Inversion of Toeplitz structured matrices using only standard equations, Linear Algebra Appl., 207 (1994), 49–70. https://doi.org/10.1016/0024-3795(94)90004-3 doi: 10.1016/0024-3795(94)90004-3
![]() |
[19] |
G. Heinig, On the reconstruction of Toeplitz matrix inverses from columns, Linear Algebra Appl., 350 (2002), 199–212. https://doi.org/10.1016/S0024-3795(02)00289-6 doi: 10.1016/S0024-3795(02)00289-6
![]() |
[20] |
L. Lerer, M. Tismenetsky, Generalized Bezoutian and the inversion problem for block matrices, Integr. Equat. Oper. Th., 9 (1986), 790–819. https://doi.org/10.1007/BF01202517 doi: 10.1007/BF01202517
![]() |
[21] |
G. Ammar, P. Gader, A variant of the Gohberg-Semencul formula involving circulant matrices, SIAM J. Matrix Anal. Appl., 12 (1991), 534–540. https://doi.org/10.1137/0612038 doi: 10.1137/0612038
![]() |
[22] |
X. G. Lv, T. Z. Huang, A note on inversion of Toeplitz matrices, Appl. Math. Lett., 20 (2007), 1189–1193. https://doi.org/10.1016/j.aml.2006.10.008 doi: 10.1016/j.aml.2006.10.008
![]() |
[23] |
Z. L. Jiang, D. D.Wang, Explicit group inverse of an innovative patterned matrix, Appl. Math. Comput., 274 (2016), 220–228. http://dx.doi.org/10.1016/j.amc.2015.11.021 doi: 10.1016/j.amc.2015.11.021
![]() |
[24] |
Z. L. Jiang, J. X. Chen, The explicit inverse of nonsingular conjugate-Toeplitz and conjugate-Hankel matrices, J. Appl. Math. Comput., 53 (2017), 1–16. http://dx.doi.org/10.1007/s12190-015-0954-y doi: 10.1007/s12190-015-0954-y
![]() |
[25] |
Z. L. Jiang, T. Y. Tam, Y. F. Wang, Inversion of conjugate-Toeplitz matrices and conjugate-Hankel matrices. Linear and Multilinear Algebra, Linear Multilinear Algebra, 65 (2017), 256–268. http://dx.doi.org/10.1080/03081087.2016.1182465 doi: 10.1080/03081087.2016.1182465
![]() |
[26] |
T. Kailath, S. Kung, M. Morf, Displacement ranks of matrices and linear equations, J. Math. Anal. Appl., 68 (1979), 395–407. http://dx.doi.org/10.1016/0022-247X(79)90124-0 doi: 10.1016/0022-247X(79)90124-0
![]() |
[27] |
I. Gohberg, V. Olshevsky, Circulants, displacements and decompositions of matrices, J. Math. Anal. Appl., 68 (1992), 730–743. http://dx.doi.org/10.1007/bf01200697 doi: 10.1007/bf01200697
![]() |
[28] |
Z. L. Jiang, T. T. Xu, Norm estimates of \omega-circulant operator matrices and isomorphic operators for \omega-circulant algebra, Sci. China Math., 59 (2016), 351–366. http://dx.doi.org/10.1007/s11425-015-5051-z doi: 10.1007/s11425-015-5051-z
![]() |
[29] |
Z. L. Jiang, Y. C. Qiao, S. D. Wang, Norm equalities and inequalities for three circulant operator matrices, Acta Math. Appl. Sin. Engl. Ser., 33 (2017), 571–590. https://doi.org/10.1007/s10114-016-5607-z doi: 10.1007/s10114-016-5607-z
![]() |
[30] |
G. Ammar, P. Gader, New decompositions of the inverse of a Toeplitz matrices, signal processing, scattering and operator theory and numerial methods, Int. Symp. MTNS-89, Birkhauser, Boston, 3 (1990), 421–428. http://dx.doi.org/10.5430/cns.v1n2p80 doi: 10.5430/cns.v1n2p80
![]() |
[31] |
P. Gader, Displacement operator based decompositions of matrices using circulants or other group matrices, Linear Algebra Appl., 139 (1990), 111–131. https://doi.org/10.1016/0024-3795(90)90392-P doi: 10.1016/0024-3795(90)90392-P
![]() |
[32] |
N. Shen, Z. L. Jiang, J. Li, On explicit determinants of the RFMLR and RLMFL circulant matrices involving certain famous numbers, WSEAS Trans. Math., 12 (2013), 42–53. http://dx.doi.org/10.1016/0044-370392-Z doi: 10.1016/0044-370392-Z
![]() |
[33] | R. A. Horn, C. R. Johnson, Matrix analysis, Cambridge university press, 1990. |
[34] | X. Y. Jiang, K. Hong, Explicit determinants of the k-Fibonacci and k-Lucas RSFPLR circulant matrix in codes, Comm. Comput. Inf. Sci., 391 (2013), 625–637. http://dx.doi.org/10.1007/978-3-642-53932-9-61 |
[35] | X. Y. Jiang, K. Hong, Exact determinants of some special circulant matrices involving four kinds of famous numbers, 2014 (2014), 1–12. http://dx.doi.org/10.1155/2014/273680 |
[36] |
X. Y. Jiang, K. Hong, Algorithms for finding inverse of two patterned matrices over \mathbb{Z}_p, Abstr. Appl. Anal., 2014 (2014), 1–6. http://dx.doi.org/10.1155/2014/840435 doi: 10.1155/2014/840435
![]() |
1. | Qi Zhang, Weiqiong Wang, Shuyu Luo, Yue Li, Repeated-root constacyclic codes of length klp over finite fields, 2025, 101, 10715797, 102542, 10.1016/j.ffa.2024.102542 | |
2. | Jinle Liu, Hongfeng Wu, A Note on Factorization and the Number of Irreducible Factors of xn − λ over Finite Fields, 2025, 13, 2227-7390, 473, 10.3390/math13030473 |