The Reed-Solomon codes are widely used to establish a reliable channel to transmit information in digital communication which has a strong error correction capability and a variety of efficient decoding algorithm. We usually use the maximum likelihood decoding algorithm (MLD) in the decoding process of Reed-Solomon codes. MLD algorithm lies in determining its error distance. Li, Wan, Hong and Wu et al obtained some results on the error distance. For the Reed-Solomon code RSq(F∗q,k), the received word u is called an ordinary word of RSq(F∗q,k) if the error distance d(u,RSq(F∗q,k))=n−deg(u(x)) with u(x) being the Lagrange interpolation polynomial of u. In this paper, we make use of the polynomial method and particularly, we use the KŚnig-Rados theorem on the number of nonzero solutions of polynomial equation over finite fields to show that if q≥4,2≤k≤q−2, then the received word u∈Fq−1q of degree q−2 is an ordinary word of RSq(F∗q,k) if and only if its Lagrange interpolation polynomial u(x) is of the form
u(x)=λq−2∑i=kaq−2−ixi+f≤k−1(x)
with a,λ∈F∗q and f≤k−1(x)∈Fq[x] being of degree at most k−1. This answers partially an open problem proposed by J.Y. Li and D.Q. Wan in [On the subset sum problem over finite fields, Finite Fields Appls. 14 (2008), 911-929].
1.
Introduction
Let Fq be the finite field of q elements with characteristic p. Let D={x1,⋯,xn} be a subset of Fq, which is called the evaluation set. The generalized Reed-Solomon code RSq(D,k) of length n and dimension k over Fq is defined as follows:
If D=F∗q, then it is called standard Reed-Solomon code, i.e.,
where α is a primitive element of Fq. We refer the above definition as the polynomial code version of the standard Reed-Solomon code. If D=Fq, then it is called the extended Reed-Solomon code. For any [n,k]q linear code C, the minimum distance d(C) is defined by
where d(⋅,⋅) denotes the Hamming distance of two words which is the number of different entries of them and w(⋅) denotes the Hamming weight of a word which is the number of its non-zero entries. Since the Reed-Solomon code is a linear code, we have
The error distance to code C of a received word u∈Fnq is defined by
Clearly, d(u,C)=0 if and only if u∈C. The most important algorithmic problem in coding theory is the maximum likelihood decoding (MLD): Given a received word u∈Fnq, find a codeword v∈C such that d(u,v)=d(u,C), then we decode u to v [4]. Therefore, it is very crucial to decide d(u,C) for the word u. When decoding the generalized Reed-Solomon code RSq(D,k), for a received word u=(u1,⋯,un)∈Fnq, we define the Lagrange interpolation polynomial u(x) of u by
i.e., u(x) is the unique polynomial of degree deg(u(x))≤n−1 such that u(xi)=ui for 1≤i≤n. For u∈Fnq, we define the degree of u(x) to be the degree of u, i.e., deg(u):=deg(u(x)). It is clear that u∈RSq(D,k) if and only if d(u,RSq(D,k))=0 if and only if deg(u)≤k−1. Equivalently, u∉RSq(D,k) if and only if d(u,RSq(D,k))≥1 if and only if k≤deg(u)≤n−1. Evidently, we have the following simple bounds due to Li and Wan [3].
Theorem 1.1. [3] Let u be a received word such that u∉RSq(D,k). Then
Let u∈Fnq. If d(u,RSq(D,k))=n−k, then the received word u is called a deep hole of RSq(D,k). In 2007, Cheng and Murray [1] conjectured that a word u is a deep hole of RSq(F∗q,k) if and only if u(x)=axk+f≤k−1(x), where u(x) is the Lagrange interpolation polynomial of the received word u and a∈F∗q. In 2012, Wu and Hong [9] disproved this conjecture by presenting a new class of deep holes for standard Reed-Solomon codes RSq(F∗q,k). In fact, let q≥4 and 2≤k≤q−2. They showed that the received word u is a deep hole if its Lagrange interpolation polynomial equals axq−2+f≤k−1(x). Later on, the main result of [9] is extended to the generalized Reed-Solomon code in [2]. Recently, some progress on deep holes of generalized projective Reed-Solomon codes are made in [10] and [11].
On the other hand, the received word u is called an ordinary word of RSq(D,k) if d(u,RSq(D,k))=n−deg(u(x)). If deg(u)=k, then the upper bound is equal to the lower bound which implies that u is a deep hole and also an ordinary word. This immediately gives (q−1)qk ordinary words. We call these trivial ordinary words. It is an interesting problem to determine all the ordinary words. In 2008, Li and Wan [3] proposed an open problem to determine all the ordinary words of the standard Reed-Solomon code. In [4], by using Weil's estimate on character sums, the following result is obtained.
Theorem 1.2. [4] Let u∈Fqq be such that k+1≤deg(u)≤q−1. Assume that q>max((deg(u))2,(deg(u)−k−1)2+ilonε) and k>(4ilonε+1)(deg(u)−k)+4ilonε+2 for some constant ilonε>0. Then u is an ordinary word of extended Reed-Solomon code RSq(Fq,k).
Furthermore, using Weil's character sum estimate and Li-Wan sieve for distinct coordinates counting, Zhu and Wan [12] showed the following result.
Theorem 1.3. [12] Let u∈Fqq be such that k+1≤deg(u)≤q−1. If there are positive constants c1 and c2 such that deg(u)−k<c1q1/2,(deg(u)−k+1)log2q<k<c2q, then u is an ordinary word of extended Reed-Solomon code RSq(Fq,k).
In [5], Li and Zhu proved the following result.
Theorem 1.4. [5] Let 3≤k+2≤q−1, and u∈Fqq be represented by polynomial u(x)=xk+2−bxk+1+cxk+v(x) with degv(x)≤k−1. If k+2=q−1 and b2=c, then u is an ordinary word of extended Reed-Solomon code RSq(Fq,k).
In this paper, we make use of a well-known result, i.e. the so-called König-Rados theorem, to find all the ordinary words of degree q−2 of standard Reed-Solomon code RSq(F∗q,k). The main result of this paper can be stated as follows.
Theorem 1.5.Let q≥4,2≤k≤q−2 and u∈Fq−1q be a received word with u(x) being its Lagrange interpolation polynomial and degu(x)=q−2. Then u is an ordinary word of RSq(F∗q,k) if and only if u(x) is of the following form
with a,λ∈F∗q and f≤k−1(x)∈Fq[x] being of degree at most k−1.
If one picks k=q−2, then the ordinary words given by Theorem 1.5 are just the trivial ones. From Theorem 1.5, the following interesting result follows immediately.
Proposition 1.6. Let q≥4,2≤k≤q−2. Then the number of ordinary words of degree q−2 of the standard Reed-Solomon code RSq(F∗q,k) is equal to (q−1)2qk.
This paper is organized as follows. First, in Section 2, we show several preliminary lemmas that are needed in the proof of Theorem 1.5. Consequently, we show Theorem 1.5 in Section 3. Finally, we present two examples to illustrate the validity of our main result.
2.
Auxiliary lemmas
In this section, our main goal is to prove several lemmas that are needed in the proof of Theorem 1.5. In what follows, we let
and
where α is a primitive element of Fq. We begin with the following lemma.
Lemma 2.1. Let u,v∈Fq−1q be two words. If u=λv+f≤k−1(F∗q), where λ∈F∗q and f≤k−1(x)∈Fq[x] is a polynomial of degree at most k−1. Then each of the following is true:
(ⅰ). We have d(u,RSq(F∗q,k))=d(v,RSq(F∗q,k)).
(ⅱ). The word u is an ordinary word of RSq(F∗q,k) if and only if the word v is an ordinary word of RSq(F∗q,k).
Proof. (ⅰ). Since RSq(F∗q,k) is a linear code, we obtain that
as desired.
(ⅱ). Since u=λv+f≤k−1(F∗q), one has degu=degv. Hence u is an ordinary word of RSq(F∗q,k) if and only if
if and only if
But part (ⅰ) tells us that d(u,RSq(F∗q,k))=d(v,RSq(F∗q,k)). So (2.1) holds if and only if
So u is ordinary holds if and only if (2.2) is true. In other words, u is ordinary if and only if v is ordinary.
This completes the proof of Lemma 2.1.
Consequently, we give another useful fact.
Lemma 2.2. Let u∈Fq−1q be a received word and u(x) be its Lagrange interpolation polynomial. Then one has
Proof. By (1.1), we have
as required.
The proof of Lemma 2.2 is complete.
The following result gives a formula on the number of nonzero solutions of polynomial equation over finite fields and is due to König and Rados (see, for instance, [6,7,8]). It is a key and important ingredient in the proof of our main result.
Lemma 2.3. (König-Rados) ([6,7,8]) Let f(x)=a0+a1x+⋯+aq−2xq−2∈Fq[x]. Then the number of nonzero solution of equation f(x)=0 in Fq is equal to q−1−rank(A), where A is the left (q−1)×(q−1) circulant matrix defined by
3.
Proof of Theorem 1.5
In this section, we use the lemmas presented in the previous section to give the proof of Theorem 1.5.
Proof of Theorem 1.5. First of all, we show the sufficiency part. Let
where a and λ∈F∗q, f≤k−1(x)∈Fq[x] is a polynomial of degree at most k−1. Define
Then u(x)=λuk(x)+f≤k−1(x). Now we pick a primitive element α of Fq and let
By Lemma 2.1, one gets that
Therefore, in order to show that
is an ordinary word, it suffices to prove that uk is an ordinary word. Equivalently, we need only to show that
since deguk(x)=q−2. This will be done in what follows.
By Lemma 2.2, we have
For any v(x)∈Pk−1, one has degv(x)≤k−1. But deguk(x)=q−2≥k. Hence
It then follows that for any v(x)∈Pk−1, one has
On the other hand, we take
Then v0(x)∈Pk−1. Furthermore, by (3.1) we have
Since
and a∈F∗q implying that
it then follows that
This infers that
from which one can derive that
So (3.4) together with (3.5) and (3.6) implies that
Hence (3.2) follows immediately from (3.3) and (3.7). So u is an ordinary word of RSq(F∗q,k). This finishes the proof of the sufficiency part.
Now we turn our attention to the proof of the necessity part. Let u be an ordinary word of RSq(F∗q,k) and degu(x)=q−2. Then by the definition of ordinary word, we have
Hence by Lemma 2.2, one has
Notice that for any v(x)∈Pk−1, one has
So there is a polynomial v0(x)∈Pk−1 such that
Write
and
Let uq−2=λ. Since degu(x)=q−2, one has λ∈F∗q. Then
with ci=λ−1ui for all integers i with k≤i≤q−2 and ci=λ−1(ui−vi) for all integers i with 0≤i≤k−1. One then deduces that
On the other hand, Lemma 2.3 yields that
where B is the left circulant matrix defined by
So from (3.8) and (3.9), we derive that rank(B)=1. Since cq−2=λ−1λ=1, one has
Assume that cq−3=0. Then B holds a nonzero minor of order 2, and so one gets that rank(B)≥2, which is impossible. Hence we must have cq−3≠0. In the following, we let cq−3=a. Then a∈F∗q. For each integer i with 1≤i≤q−1, let ri denote the i-th row of the matrix B. Then rank(ri)=1 for each integer i with 1≤i≤q−1. Since rank(B) = 1, there exists an element a∈F∗q such that r1=ar2. Then we deduce that cq−2=ac0 and ck=ack+1 for 0≤k≤q−3. Since cq−2=1, one has a=cq−3,c0=a−1=aq−2 and ck=a−1ck−1 for 1≤k≤q−2. It then follows that ck=aq−2−k for 0≤k≤q−2. This implies that
Therefore
where
So the necessity part is proved.
This concludes the proof of Theorem 1.5
4.
Examples and final remarks
In this last section, we supply two examples to demonstrate the validity of Theorem 1.5.
Example 4.1. Let q=7,n=q−1=6,k=3. Putting α=3 gives us the standard Reed-Solomon code
Using the MATLAB 2011a programming, we search for the ordinary word and find out all the ordinary words of degree q−2=5 of standard Reed-Solomon code RS7(F∗7,3) that are listed in Table 1. By (1.2), we can get the Lagrange interpolation polynomial u(x) of the ordinary word u of degree 5 of RS7(F∗7,3) listed also in Table 1. This coincides with Theorem 1.5.
Suppose that u is an ordinary word of degree 5 of RS7(F∗7,3). Then
On the other hand, one has
So it is sufficient to find a codeword v in RS7(F∗7,3) such that d(u,v)=1. For the received ordinary word u=λ(3,1,0,6,0,4)+f of degree 5, by (1.2) we compute and get that u(x)=λ5∑i=3xi+f(x). Furthermore, one can search and find the word v=λ(4,1,0,6,0,4)+f∈RS7(F∗7,3) such that d(u,v)=1. For the other ordinary words u of degree 5, one can also find the corresponding codewords v such that d(u,v)=1. We can easily compute the Lagrange interpolation polynomial v(x) of v also listed in Table 1.
Example 4.2. Let q=11,n=q−1=10,k=5. Putting α=2 gives us the standard Reed-Solomon code
Using the MATLAB 2011a programming, we search for the ordinary word and find out all the ordinary words of degree q−2=9 of standard Reed-Solomon code RS11(F∗11,5) that are listed in Table 2. By (1.2), one can easily calculate the Lagrange interpolation polynomial u(x) of the ordinary word u of degree 9 of RS11(F∗11,5) listed also in Table 2. This coincides with Theorem 1.5.
Suppose that u is an ordinary word of degree 9 of RS11(F∗11,5). Then
On the other hand, one has
So it is sufficient to find a codeword v in RS11(F∗11,5) such that d(u,v)=1. For the received ordinary word u=λ(5,2,0,5,0,10,0,4,0,7)+f of degree 9, by (1.2) we compute and get that u(x)=λ9∑i=5xi+f(x). Furthermore, one can search and find the codeword v=λ(6,2,0,5,0,10,0,4,0,7)+f∈RS11(F∗11,5) such that d(u,v)=1. For the other ordinary words u of degree 9, one can also find the corresponding codewords v such that d(u,v)=1. It is easy to compute the Lagrange interpolation polynomial v(x) of v that is listed in Table 2.
Remark 4.3. In this paper, we determine all the ordinary words of maximal degree q−2 of the standard Reed-Solomon code RSq(F∗q,k). In the close future, we will explore the ordinary words of degree no more than q−3 of the standard Reed-Solomon code RSq(F∗q,k).
Acknowledgments
The authors would like to thank the anonymous referees for their careful readings of the manuscript and helpful comments and suggestions that improved the presentation of the paper. This work was supported partially by National Science Foundation of China Grant # 11771304, by the Fundamental Research Funds for the Central Universities and by Foundation of Sichuan Tourism University under Grant # 19SCTUZZ01.
Conflict of interest
We declare that we have no conflict of interest.