Abstract
Symbol-pair codes are proposed to guard against pair-errors in symbol-pair read channels. The minimum symbol-pair distance plays a vital role in determining the error-correcting capability and the constructions of symbol-pair codes with largest possible minimum symbol-pair distance is of great importance. Maximum distance separable (MDS) symbol-pair codes are optimal in the sense that such codes can acheive the Singleton bound. In this paper, for length 5p, two new classes of MDS symbol-pair codes with minimum symbol-pair distance seven or eight are constructed by utilizing repeated-root cyclic codes over \({\mathbb {F}}_{p}\), where p is a prime. In addition, we derive a class of MDS symbol-pair codes with minimum symbol-pair distance seven and length 4p.
Similar content being viewed by others
References
Cassuto Y., Blaum M.: Codes for symbol-pair read channels. In: Proceedings of IEEE International Symposium Information Theory (ISIT), pp. 988–992 (2010).
Cassuto Y., Blaum M.: Codes for symbol-pair read channels. IEEE Trans. Inf. Theory 57(12), 8011–8020 (2011).
Cassuto Y., Litsyn S.: Symbol-pair codes: algebraic constructions and asymptotic bounds. In: Proceedings of IEEE International Symposium Information Theory (ISIT), pp. 2348–2352 (2011).
Castagnoli G., Massey J.L., Schoeller P.A., von Seemann N.: On repeated-root cyclic codes. IEEE Trans. Inf. Theory 37(2), 337–342 (1991).
Chee Y.M., Ji L., Kiah H.M., Wang C., Yin J.: Maximum distance separable codes for symbol-pair read channels. IEEE Trans. Inf. Theory 59(11), 7259–7267 (2013).
Chee Y.M., Kiah H.M., Wang C.: Maximum distance separable symbol-pair codes. In: Proceedings of IEEE International Symposium Information Theory (ISIT), pp. 2886–2890 (2012).
Chen B., Lin L., Liu H.: Constacyclic symbol-pair codes: lower bounds and optimal constructions. IEEE Trans. Inf. Theory 63(12), 7661–7666 (2017).
Ding B., Ge G., Zhang J., Zhang T., Zhang Y.: New constructions of MDS symbol-pair codes. Des. Codes Cryptogr. 86(4), 841–859 (2018).
Ding B., Zhang T., Ge G.: Maximum distance separable codes for \(b\)-symbol read channels. Finite Fields Appl. 49(1), 180–197 (2018).
Dinh H.Q., Nguyen B.T., Singh A.K., Sriboonchitta S.: Hamming and symbol-pair distances of repeated-root constacyclic codes of prime power lengths over \({\mathbb{F}}_{p^m}+u\,{\mathbb{F}}_{p^{m}}\). IEEE Commun. Letters 22(12), 2400–2403 (2018).
Dinh H.Q., Nguyen B.T., Singh A.K., Sriboonchitta S.: On the symbol-pair distance of repeated-root constacyclic codes of prime power lengths. IEEE Trans. Inf. Theory 64(4), 2417–2430 (2018).
Dinh H.Q., Nguyen B.T., Sriboonchitta S.: MDS symbol-pair cyclic codes of length \(2p^s\) over \({\mathbb{F}}_{p^m}\). IEEE Trans. Inf. Theory 66(1), 240–262 (2020).
Dinh H.Q., Wang X., Liu H., Sriboonchitta S.: On the symbol-pair distances of repeated-root constacyclic codes of length \(2p^s\). Discret Math. 342(11), 3062–3078 (2019).
Elishco O., Gabrys R., Yaakobi E.: Bounds and constructions of codes over symbol-pair read channels. IEEE Trans. Inf. Theory 66(3), 1385–1395 (2020).
Horii S., Matsushima T., Hirasawa S.: Linear programming decoding of binary linear codes for symbol-pair read channels. In: Proceedings of IEEE International Symposium Information Theory (ISIT), pp. 1944–1948 (2016).
Huffman W.C., Pless V.: Fundamentals of Error-Correcting Codes. Cambridge University Press, Cambridge, U.K. (2003).
Kai X., Zhu S., Li P.: A construction of new MDS symbol-pair codes. IEEE Trans. Inf. Theory 61(11), 5828–5834 (2015).
Kai X., Zhu S., Zhao Y., Luo H., Chen Z.: New MDS symbol-pair codes from repeated-root codes. IEEE Commun. Lett. 22(3), 462–465 (2018).
Li S., Ge G.: Constructions of maximum distance separable symbol-pair codes using cyclic and constacyclic codes. Des. Codes Cryptogr. 84(3), 359–372 (2017).
Liu S., Xing C., Yuan C.: List decodability of symbol-pair codes. IEEE Trans. Inf. Theory 65(8), 4815–4821 (2019).
Ma J., Luo J.: MDS symbol-pair codes from repeated-root cyclic codes. Des. Codes Cryptogr. 90(1), 121–137 (2022).
Ma J., Luo J.: On symbol-pair weight distribution of MDS codes and simplex codes over finite fields. Cryptogr. Commun. 13(1), 101–115 (2021).
MacWilliams F.J., Sloane N.J.A.: The Theory of Error-Correcting Codes. North. Holland, Amsterdam (1977).
Massey J.L., Costello D.J., Justesen J.: Polynomial weights and code constructions. IEEE Trans. Inf. Theory IT-19(1), 101–110 (1973).
Morii M., Hirotomo M., Takita M.: Error-trapping decoding for cyclic codes over symbol-pair read channels. In: Proceedings of IEEE International Symposium Information Theory (ISIT), pp. 681–685 (2016).
Sun Z., Zhu S., Wang L.: The symbol-pair distance distribution of a class of repeated-root cyclic codes over \({\mathbb{F}}_{p^m}\). Cryptogr. Commun. 10(4), 643–653 (2018).
Takita M., Hirotomo M., Morii M.: Syndrome decoding of symbol-pair codes. IEICE Trans. A 98(12), 2423–2428 (2015).
Yaakobi E., Bruck J., Siegel P. H.: Decoding of cyclic codes over symbol-pair read channels. In: Proceedings of IEEE International Symposium Information Theory (ISIT), pp. 2891–2895 (2012).
Acknowledgements
This work was supported by National Natural Science Foundation of China (Nos. 12171191, 11871025), in part by Hubei Provincial Science and Technology Innovation Base (Platform) Special Project (No. 2020DFH002) and Application Foundation Frontier Project of Wuhan Science and Technology Bureau (No. 2020010601012189).
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by G. Ge.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix Proof of Theorem 2:
Appendix Proof of Theorem 2:
Recall that \({\mathcal {C}}\) is a repeated-root cyclic code of length 5p over \({\mathbb {F}}_{p}\) with the generator polynomial
where \(\beta \) is a primitive 5-th root of unity in \({\mathbb {F}}_{p}\). By Lemma 3, one can derive that the parameter of \({\mathcal {C}}\) is \(\left[ 5p,\,5p-6,\,4\right] \). Since \({\mathcal {C}}\) is not MDS, Lemma 4 yields that \(d_{p}({\mathcal {C}})\ge 6\). Similar as Theorem 1, one can derive that there does not exist a codeword c(x) in \({\mathcal {C}}\) with \((w_H(c(x)), \,w_p(c(x)))=(5,\,6)\) or \((6,\,7)\). To prove that \({\mathcal {C}}\) is an MDS \(\left( 5p,\,8\right) _{p}\) symbol-pair code, it suffices to determine that there does not exist a codeword in \({\mathcal {C}}\) with \((w_H(c(x)), \,w_p(c(x)))=(4,\,6)\), \((4,\,7)\) or \((5,\,7)\).
Case I \((w_H(c(x)), \,w_p(c(x)))=(4,\,6)\). Since \({\mathcal {C}}\) is the subcode of the code occurred in Theorem 1 and the proof of Theorem 1 indicates that there does not exist a codeword c(x) in \({\mathcal {C}}\) with \((w_H(c(x)), \,w_p(c(x)))=(4,\,6)\) unless \(p=41\). Now it is sufficient to show that for \(p=41\), there does not exist a codeword c(x) in \({\mathcal {C}}\) with \((w_H(c(x)), \,w_p(c(x)))=(4,\,6)\). More precisely, we just need to consider Case II in Theorem 1. There are two subcases to be discussed:
-
Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{2}+a_{3}\,x^{5i+3}\) with \(1\le i\le p-1\) and \(a_{1},\,a_{2},\,a_{3}\in {\mathbb {F}}_p^{*}\). Notice that \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) and Lemma 6 induces
$$\begin{aligned} a_{1}=-\frac{\beta ^2+\beta +1}{\beta ^2},\quad a_{2}=\frac{\beta ^2+\beta +1}{\beta ^3},\quad a_{3}=-\frac{1}{\beta ^3}. \end{aligned}$$(17)It follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+\left( 5i+3\right) a_{3}=0,\\ a_{1}+2\,a_{2}\,\beta ^2+\left( 5i+3\right) a_{3}\,\beta ^4=0\\ \end{array} \right. \end{aligned}$$which yields
$$\begin{aligned} a_1\left( \beta ^4-1\right) +2\,a_2\left( \beta ^4-\beta ^2\right) =0. \end{aligned}$$Combining with (17), one can get \(\left( \beta -1\right) ^2=0\), a contradiction.
-
For the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{2}+a_{3}\,x^{5i+4}\) with \(0\le i\le p-2\) and \(a_{1},\,a_{2},\,a_{3}\in {\mathbb {F}}_p^{*}\), by \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) and Lemma 6, one can obtain that
$$\begin{aligned} a_{1}=-\frac{1}{\beta },\quad a_{2}=-\frac{\beta }{\beta +1},\quad a_{3}=\frac{1}{\beta \left( \beta +1\right) }. \end{aligned}$$(18)On the other hand, \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+\left( 5i+4\right) a_{3}=0,\\ a_{1}+2\,a_{2}\,\beta ^2+\left( 5i+4\right) a_{3}\,\beta =0\\ \end{array} \right. \end{aligned}$$which induces
$$\begin{aligned} a_1\left( \beta -1\right) =2\,a_2\left( \beta ^2-\beta \right) . \end{aligned}$$Together with (18), one can immediately obtain that
$$\begin{aligned} 2\,\beta ^3=\beta +1. \end{aligned}$$This leads to
$$\begin{aligned} \left( \beta -1\right) \left( 2\,\beta ^2+2\,\beta +1\right) =0. \end{aligned}$$The fact \(\beta \) is a primitive 5-th root of unity implies that \(2\beta ^2+2\beta +1=0\) and then one has
$$\begin{aligned} \beta ^2+\beta =-\left( \beta ^2+\beta +1\right) =\beta ^4+\beta ^3 \end{aligned}$$which is impossible.
Case II \((w_H(c(x)), \,w_p(c(x)))=(4,\,7)\). For this case, Lemma 1 implies that the cyclic shift of c(x) must have the form
Assume that \(c(x)=\left( x^{5}-1\right) ^{t}v(x)\), where \(0\le t\le p-1\), \(\left( x^{5}-1\right) \not \mid v(x)\) and
Recall that \(N_{v}=w_{H}\left( v(x)\,\mathrm{mod}\left( x^{5}-1\right) \right) \). Then by Lemma 5, one can deduce that
If \(\left( N_{v},\,t\right) =\left( 1,\,3\right) \), then it is easily seen that the symbol-pair weight of c(x) is greater than 7.
If \(\left( N_{v},\,t\right) =\left( 2,\,1\right) \), then there are three subcases to be discussed:
(1) For the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{5i}+a_{3}\,x^{5j}\) with \(1\le i<j \le p-1\) and \(a_{1},\,a_{2},\,a_{3}\in {\mathbb {F}}_p^{*}\), it can be verified that
since \(c\left( \beta \right) =c\left( \beta ^2\right) =0\). Then one can obtain that \(a_{1}=0\), a contradiction.
(2) For the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{5i+1}+a_{3}\,x^{5j+1}\) with \(1\le i<j \le p-1\) and \(a_{1},\,a_{2},\,a_{3}\in {\mathbb {F}}_p^{*}\), by \(c\left( 1\right) =c\left( \beta \right) =0\), one can get
This implies that \(\beta =1\), which is impossible.
(3) For the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{5i}+a_{3}\,x^{5j+1}\) with \(1\le i<j \le p-1\) and \(a_{1},\,a_{2},\,a_{3}\in {\mathbb {F}}_p^{*}\), it follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
This leads to \(\beta ^3=1\), a contradiction.
If \(\left( N_{v},\,t\right) =\left( 4,\,0\right) \), then there are also three subcases to be considered:
(1) For the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{5i+2}+a_{3}\,x^{5j+3}\) with \(1\le i<j\le p-1\) and \(a_{1},\,a_{2},\,a_{3}\in {\mathbb {F}}_p^{*}\), by Lemma 6 and \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\), one can derive that
It follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
which indicates
Together with (19), one can immediately obtain that
By substituting the value of \(\beta ^2+1\) in the first equality into the second equality of (20), we can get
which yields \(i=j\) due to \(p\not \mid (5i+2)\). This contradicts with \(i<j\).
(2) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{5i+2}+a_{3}\,x^{5j+4}\) with \(1\le i\le j\le p-2\) and \(a_{1},\,a_{2},\,a_{3}\in {\mathbb {F}}_p^{*}\). The fact \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) and Lemma 6 leads to
It follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
By substituting (21), one can immediately derive that
This leads to \(\left( 5i+2\right) ^2=5j+4\). Since it can be verified that \(p\not \mid \left( 5i+2\right) \), it follows from \(c^{\left( 2\right) }\left( 1\right) =0\) that
Then (21) and \(c^{\left( 1\right) }\left( 1\right) =0\) indicates that
Let \(t=5i+2\). Then one has \(\beta +1=t\beta ^3\) and \(\beta ^2=t\left( t+1\right) \) due to the first equality of (22) and (23). It follows from (24) that
which implies \(\beta +1=-t^3\). Combining with \(\beta +1=t\beta ^3\), we have \(\beta ^3=-t^2\). Since \(\beta \) is a primitive 5-th root of unity, one can derive
It follows from \(t\left( t+1\right) =\beta ^2\ne 0\) that \(t^3-t^2+1=0\). Then we obtain
which yields \(\beta ^2-1=0\), a contradiction.
(3) For the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{5i+3}+a_{3}\,x^{5j+4}\) with \(0\le i<j\le p-2\) and \(a_{1},\,a_{2},\,a_{3}\in {\mathbb {F}}_p^{*}\), it follows from \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) and Lemma 6 that
Since \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\), one can immediately get
Together with (25), one can conclude that
which indicates
It follows that \(5(i-j)=0\), a contradiction.
Case III \((w_H(c(x)), \,w_p(c(x)))=(5,\,7)\). In this case, we can assume that c(x) is of the form
where \(\mathbf {a}\), \(\mathbf {b}\) are row vectors with all entries of \(\mathbf {a}\), \(\mathbf {b}\) being nonzero. Then its certain cyclic shift must have the form
or
-
For \(\left( \star ,\,\star ,\,\star ,\,\star ,\,\mathbf {0},\,\star ,\,\mathbf {0}\right) \), there are five subcases to be considered:
(1) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^{2}+a_{3}\,x^3+a_{4}\,x^{5i}\) with \(1\le i\le p-1\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). It can be verified that
$$\begin{aligned} \left\{ \begin{array}{l} 1+a_{1}+a_{2}+a_{3}+a_{4}=0,\\ 1+a_{1}\,\beta +a_{2}\,\beta ^2+a_{3}\,\beta ^3+a_{4}=0,\\ 1+a_{1}\,\beta ^2+a_{2}\,\beta ^4+a_{3}\,\beta +a_{4}=0\\ \end{array} \right. \end{aligned}$$since \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\). Then one can derive that \(p\not \mid \left( a_{4}+1\right) \). By Lemma 6, one can obtain
$$\begin{aligned} a_{1}=-\frac{\beta ^2+\beta +1}{\beta ^2}(a_{4}+1), a_{2}=\frac{\beta ^2+\beta +1}{\beta ^3}(a_{4}+1), a_{3}=-\frac{1}{\beta ^3}(a_{4}+1). \end{aligned}$$(26)It follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+3\,a_{3}+5i\,a_{4}=0,\\ a_{1}+2\,a_{2}\,\beta ^2+3\,a_{3}\,\beta ^4+5i\,a_{4}\,\beta ^3=0\\ \end{array} \right. \end{aligned}$$which indicates
$$\begin{aligned} \left( \beta ^3-1\right) a_{1}+2\left( \beta ^3-\beta ^2\right) a_{2} +3\left( \beta ^3-\beta ^4\right) a_{3}=0. \end{aligned}$$Combining with (26), one can derive that
$$\begin{aligned} -\left( \beta ^3-1\right) \beta \left( \beta ^2+\beta +1\right) +2\beta ^2\left( \beta -1\right) \left( \beta ^2+\beta +1\right) +3\beta ^3\left( \beta -1\right) =0. \end{aligned}$$Since \(\beta \) is a primitive 5-th root of unity, by expanding the above equality, one can get \(\beta ^2+3\beta +1=0\). This is contradictory with the inequality (3) in Lemma 6.
(2) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^3+a_{4}\,x^{5i+1}\) with \(1\le i\le p-1\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). It follows from \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) and Lemma 6 that
$$\begin{aligned} a_{1}+a_{4}=-\frac{\beta ^2+\beta +1}{\beta ^2},\quad a_{2}=\frac{\beta ^2+\beta +1}{\beta ^3},\quad a_{3}=-\frac{1}{\beta ^{3}}. \end{aligned}$$(27)Then \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) induces that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+3\,a_{3}+\left( 5i+1\right) a_{4}=0,\\ a_{1}+2\,a_{2}\,\beta ^2+3\,a_{3}\,\beta ^4+\left( 5i+1\right) a_{4}=0.\\ \end{array} \right. \end{aligned}$$This leads to
$$\begin{aligned} 2\left( \beta ^2-1\right) a_{2}+3\left( \beta ^4-1\right) a_{3}=0. \end{aligned}$$Together with (27), one can immediately get \(\left( \beta -1\right) ^2=0\), which is impossible.
(3) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^3+a_{4}\,x^{5i+2}\) with \(1\le i\le p-1\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). The fact \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) and Lemma 6 induces
$$\begin{aligned} a_{1}=-\frac{\beta ^2+\beta +1}{\beta ^2},\quad a_{2}+a_{4}=\frac{\beta ^2+\beta +1}{\beta ^3},\quad a_{3}=-\frac{1}{\beta ^3}. \end{aligned}$$(28)It follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+3\,a_{3}+\left( 5i+2\right) a_{4}=0,\\ a_{1}+2\,a_{2}\,\beta ^2+3\,a_{3}\,\beta ^4+\left( 5i+2\right) a_{4}\,\beta ^2=0\\ \end{array} \right. \end{aligned}$$which implies
$$\begin{aligned} \left( \beta ^2-1\right) a_{1}+3\left( \beta ^2-\beta ^4\right) a_{3}=0. \end{aligned}$$By substituting (28) into the above equality, we have \(\left( \beta -1\right) ^2=0\), a contradiction.
(4) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^3+a_{4}\,x^{5i+3}\) with \(1\le i\le p-1\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). By \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) and Lemma 6, one has
$$\begin{aligned} a_{1}=-\frac{\beta ^2+\beta +1}{\beta ^2},\quad a_{2}=\frac{\beta ^2+\beta +1}{\beta ^3},\quad a_{3}+a_{4}=-\frac{1}{\beta ^3}. \end{aligned}$$(29)It follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+3\,a_{3}+\left( 5i+3\right) a_{4}=0,\\[2mm] a_{1}+2\,a_{2}\,\beta ^2+3\,a_{3}\,\beta ^4+\left( 5i+3\right) a_{4}\,\beta ^4=0.\\ \end{array} \right. \end{aligned}$$This yields
$$\begin{aligned} \left( \beta ^4-1\right) a_{1}+2\left( \beta ^4-\beta ^2\right) a_{2}=0. \end{aligned}$$Combining with (29), one can derive that \(\left( \beta -1\right) ^2=0\), which is impossible.
(5) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^3+a_{4}\,x^{5i+4}\) with \(1\le i\le p-2\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). It can be verified that
$$\begin{aligned} \left\{ \begin{array}{l} 1+a_{1}+a_{2}+a_{3}+a_{4}=0,\\ 1+a_{1}\,\beta +a_{2}\,\beta ^2+a_{3}\,\beta ^3+a_{4}\,\beta ^4=0,\\ 1+a_{1}\,\beta ^2+a_{2}\,\beta ^4+a_{3}\,\beta +a_{4}\,\beta ^3=0\\ \end{array} \right. \end{aligned}$$since \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\). Then one can obtain that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}=-\beta ^3\,a_{4}+\beta ^2+\beta ,\\ a_{2}=-\left( \beta ^4+1\right) a_{4}-\beta -1,\\ a_{3}=-\left( \beta ^2+\beta +1\right) a_{4}-\beta ^2.\\ \end{array} \right. \end{aligned}$$(30)It follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+3\,a_{3}+\left( 5i+4\right) a_{4}=0,\\[2mm] a_{1}+2\,a_{2}\,\beta ^2+3\,a_{3}\,\beta ^4+\left( 5i+4\right) a_{4}\,\beta =0\\ \end{array} \right. \end{aligned}$$which implies
$$\begin{aligned} \left( \beta -1\right) a_{1}+2\left( \beta -\beta ^2\right) a_{2} +3\left( \beta -\beta ^4\right) a_{3}=0. \end{aligned}$$This is equivalent to
$$\begin{aligned} a_1-2\,\beta \,a_2-3\,\beta \left( \beta ^2+\beta +1\right) a_3=0. \end{aligned}$$Together with (30), one can immediately have
$$\begin{aligned} (-\beta ^3+2\beta (\beta ^4+1)+3\,\beta (\beta ^2+\beta +1)^2)a_4+\beta ^2+\beta +2\,\beta (\beta +1) +3\,\beta ^3(\beta ^2+\beta +1)=0. \end{aligned}$$Then we get that
$$\begin{aligned} -\beta ^3+2\beta \left( \beta ^4+1\right) +3\,\beta \left( \beta ^2+\beta +1\right) ^2=0 \end{aligned}$$due to \(\beta ^4+\beta ^3+\beta ^2+\beta +1=0\) and \(a_4\in {\mathbb {F}}_{p}^{*}\). By a straightforward computation, one has \(\beta ^2+3\beta +1=0\). This contradicts with the inequality (3) in Lemma 6.
-
For \(\left( \star ,\,\star ,\,\star ,\,\mathbf {0},\,\star ,\,\star ,\,\mathbf {0}\right) \), there are also five subcases to be considered:
(1) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^{5i}+a_{4}\,x^{5i+1}\) with \(1\le i\le p-1\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). It follows from \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) that
$$\begin{aligned} \left\{ \begin{array}{l} 1+a_{1}+a_{2}+a_{3}+a_{4}=0,\\ 1+a_{1}\,\beta +a_{2}\,\beta ^2+a_{3}+a_{4}\,\beta =0,\\ 1+a_{1}\,\beta ^2+a_{2}\,\beta ^4+a_{3}+a_{4}\,\beta ^2=0\\ \end{array} \right. \end{aligned}$$which implies
$$\begin{aligned} \left\{ \begin{array}{l} \left( a_{1}+a_{4}\right) \left( \beta -1\right) +a_{2}\left( \beta ^2-1\right) =0,\\ \left( a_{1}+a_{4}\right) \left( \beta ^2-\beta \right) +a_{2}\left( \beta ^4-\beta ^2\right) =0.\\ \end{array} \right. \end{aligned}$$This indicates that \(\beta \left( \beta ^2-1\right) a_{2}=\left( \beta ^4-\beta ^2\right) a_{2}\). Hence \(\beta =1\), a contradiction.
(2) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^{5i+1}+a_{4}\,x^{5i+2}\) with \(1\le i\le p-1\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). It can be verified that
$$\begin{aligned} \left\{ \begin{array}{l} 1+a_{1}+a_{2}+a_{3}+a_{4}=0,\\ 1+a_{1}\,\beta +a_{2}\,\beta ^2+a_{3}\,\beta +a_{4}\,\beta ^2=0,\\ 1+a_{1}\,\beta ^2+a_{2}\,\beta ^4+a_{3}\,\beta ^2+a_{4}\,\beta ^4=0\\ \end{array} \right. \end{aligned}$$since \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\). Then one can derive that
$$\begin{aligned} \left\{ \begin{array}{l} \left( a_{2}+a_{4}\right) \left( \beta ^2-\beta \right) =\beta -1,\\ \left( a_{2}+a_{4}\right) \left( \beta ^4-\beta ^3\right) =\beta -1.\\ \end{array} \right. \end{aligned}$$It follows that \(\beta ^3=\beta \), which is impossible.
(3) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^{5i+2}+a_{4}\,x^{5i+3}\) with \(1\le i\le p-1\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). The fact \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) and Lemma 6 induces that
$$\begin{aligned} a_{1}=-\frac{\beta ^2+\beta +1}{\beta ^2},\quad a_{2}+a_{3}=\frac{\beta ^2+\beta +1}{\beta ^3},\quad a_{4}=-\frac{1}{\beta ^3}. \end{aligned}$$(31)It follows from \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+\left( 5i+2\right) a_{3}+\left( 5i+3\right) a_{4}=0,\\ a_{1}+2\,a_{2}\,\beta ^2+\left( 5i+2\right) a_{3}\,\beta ^2 +\left( 5i+3\right) a_{4}\,\beta ^4=0.\\ \end{array} \right. \end{aligned}$$This yields
$$\begin{aligned} \left( \beta ^2-1\right) a_{1}+\left( 5i+3\right) \left( \beta ^2-\beta ^4\right) a_{4}=0. \end{aligned}$$By substituting (31), one can deduce that
$$\begin{aligned} \beta ^2-\left( 5i+2\right) \beta +1=0. \end{aligned}$$Let \(t=5i+2\). Then \(\beta ^2=t\beta -1\) and
$$\begin{aligned} \beta ^4+\beta ^3+\beta ^2+\beta +1=(t\beta -1)(t^2+t-1)=0. \end{aligned}$$It follows that \(t^2+t=1\). By \(c^{\left( 2\right) }\left( 1\right) =0\) and (31), we get
$$\begin{aligned} 5i\left( t+1\right) a_{3}=\left( t+2\right) \beta +1. \end{aligned}$$The fact \(c^{\left( 1\right) }\left( 1\right) =0\) indicates \(5i\,a_{3}=\left( 2-t\right) \left( \beta +1\right) \). Hence
$$\begin{aligned} \left( t+2\right) \beta +1=\left( t+1\right) \left( 2-t\right) \left( \beta +1\right) . \end{aligned}$$This leads to \(t^2\,\beta -2\,t=0\) due to \(t^2+t=1\). It follows from \(t\ne 0\) that \(t\beta =2\) and \(\beta ^2=t\beta -1=1\), a contradiction.
(4) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^{5i+3}+a_{4}\,x^{5i+4}\) with \(1\le i\le p-2\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). It can be checked that
$$\begin{aligned} \left\{ \begin{array}{l} 1+a_{1}+a_{2}+a_{3}+a_{4}=0,\\ 1+a_{1}\,\beta +a_{2}\,\beta ^2+a_{3}\,\beta ^3+a_{4}\,\beta ^4=0,\\ 1+a_{1}\,\beta ^2+a_{2}\,\beta ^4+a_{3}\,\beta +a_{4}\,\beta ^3=0\\ \end{array} \right. \end{aligned}$$since \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\). Then one can derive that
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}=-\beta ^3\,a_{4}+\beta ^2+\beta ,\\ a_{2}=-\left( \beta ^4+1\right) a_{4}-\beta -1,\\ a_{3}=-\left( \beta ^2+\beta +1\right) a_{4}-\beta ^2.\\ \end{array} \right. \end{aligned}$$(32)Let \(t=5i+2\). By \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) and (32), we have
$$\begin{aligned} \left\{ \begin{array}{l} t\beta ^2+\beta +2=\left( (t-1)\beta ^4+t\beta ^3+t\right) a_{4},\\ 2\beta ^2+\beta +t=\left( t\beta ^2+(t-1)\beta +t\right) a_{4}.\\ \end{array} \right. \end{aligned}$$(33)Then
$$\begin{aligned} (t\beta ^2+\beta +2)\left( t\beta ^2+(t-1)\beta +t\right) =(2\beta ^2+\beta +t)\left( (t-1)\beta ^4+t\beta ^3+t\right) \end{aligned}$$which implies
$$\begin{aligned} (t^2+t-1)(\beta ^2-1)=0. \end{aligned}$$Thus \(t^2+t=1\). It follows from \(c^{\left( 2\right) }\left( 1\right) =0\) that \(2\,a_{2}+a_{3}+\left( 2t+3\right) a_{4}=0\). Together with (32), one can immediately get
$$\begin{aligned} \left( -\beta ^4+\beta ^3+2t+1\right) a_{4}=\beta ^2+2\beta +2. \end{aligned}$$Combining with the second equality in (33), we can obtain
$$\begin{aligned} \left( -\beta ^4+\beta ^3+2t+1\right) \left( 2\beta ^2+\beta +t\right) =\left( \beta ^2+2\beta +2\right) \left( t\beta ^2+(t-1)\beta +t\right) . \end{aligned}$$By expanding the above equality, one can deduce
$$\begin{aligned} \left( \beta ^2-1\right) t+3\beta ^2+2=0 \end{aligned}$$which yields \(t=\frac{3\beta ^2+2}{1-\beta ^2}\). The fact \(t^2+t-1=0\) induces
$$\begin{aligned} \left( \frac{3\beta ^2+2}{1-\beta ^2}\right) ^2+\frac{3\beta ^2+2}{1-\beta ^2}-1=0 \end{aligned}$$which is equivalent to
$$\begin{aligned} \left( 3\beta ^2+2\right) ^2+\left( 3\beta ^2+2\right) \left( 1-\beta ^2\right) -\left( 1-\beta ^2\right) ^2=0. \end{aligned}$$It follows that
$$\begin{aligned} \beta ^4+3\,\beta ^2+1=0 \end{aligned}$$which indicates
$$\begin{aligned} 2\,\beta ^2-\beta ^3-\beta =0 \end{aligned}$$due to \(\beta ^4+\beta ^3+\beta ^2+\beta +1=0\). Hence \(\beta (\beta -1)^2=0\), which is impossible.
(5) Consider the subcase \(c(x)=1+a_{1}\,x+a_{2}\,x^2+a_{3}\,x^{5i+4}+a_{4}\,x^{5i+5}\) with \(0\le i\le p-2\) and \(a_{1},\,a_{2},\,a_{3},\,a_{4}\in {\mathbb {F}}_{p}^{*}\). It follows from \(c\left( 1\right) =c\left( \beta \right) =c\left( \beta ^2\right) =0\) that \(p\not \mid \left( a_{4}+1\right) \) and
$$\begin{aligned} a_{1}=-\frac{1}{\beta }\left( a_{4}+1\right) ,\quad a_{2}=-\frac{\beta }{\beta +1}\left( a_{4}+1\right) ,\quad a_{3}=\frac{1}{\beta (\beta +1)}\left( a_{4}+1\right) \end{aligned}$$(34)due to Lemma 6. The fact \(c^{\left( 1\right) }\left( 1\right) =c^{\left( 1\right) }\left( \beta ^2\right) =0\) leads to
$$\begin{aligned} \left\{ \begin{array}{l} a_{1}+2\,a_{2}+\left( 5i+4\right) a_{3}+\left( 5i+5\right) a_{4}=0,\\[2mm] a_{1}+2\,a_{2}\,\beta ^2+\left( 5i+4\right) a_{3}\,\beta +\left( 5i+5\right) a_{4}\,\beta ^3=0\\ \end{array} \right. \end{aligned}$$which implies
$$\begin{aligned} \left\{ \begin{array}{l} \left( \beta ^3-1\right) a_{1}+2\left( \beta ^3-\beta ^2\right) a_{2} +\left( 5i+4\right) \left( \beta ^3-\beta \right) a_{3}=0,\\[2mm] 2\left( \beta +1\right) a_{2}+\left( 5i+4\right) a_{3}+\left( 5i+5\right) \left( \beta ^2+\beta +1\right) a_{4}=0.\\ \end{array} \right. \end{aligned}$$By substituting (34), one can obtain that
$$\begin{aligned} \beta ^4+\left( 5i+3\right) \beta ^3-\left( 5i+3\right) \beta -1=0 \end{aligned}$$(35)and
$$\begin{aligned} \left( -2\beta ^2\left( \beta +1\right) +5i+4\right) \left( a_{4}+1\right) +\left( 5i+5\right) \beta \left( \beta +1\right) \left( \beta ^2+\beta +1\right) a_4=0. \end{aligned}$$(36)Let \(t=5i+3\). It follows from (35) that
$$\begin{aligned} \beta ^4-1+t\left( \beta ^3-\beta \right) =\left( \beta ^2-1\right) (\beta ^2+1+t\,\beta )=0 \end{aligned}$$which yields \(\beta ^2=-t\beta -1\). Then we have
$$\begin{aligned} 0=\beta ^4+\beta ^3+\beta ^2+\beta +1=-\left( t^2-t-1\right) \beta ^2 \end{aligned}$$which indicates \(t^2=t+1\) due to \(\beta ^2\ne 0\). It can be verified that
$$\begin{aligned} \begin{aligned}&-2\beta ^2\left( \beta +1\right) +5i+4+\left( 5i+5\right) \beta \left( \beta +1\right) \left( \beta ^2+\beta +1\right) \\&\quad =-2\beta ^3-2\beta ^2+t+1-\left( t+2\right) \left( \beta ^4+\beta +2\right) \\&\quad =-2t\left( \beta +1\right) +2\left( t\beta +1\right) +\left( t+2\right) \left( \beta +t\right) -\left( t+2\right) \beta -t-3=0. \end{aligned} \end{aligned}$$Hence (36) and \(a_4\in {\mathbb {F}}_{p}^{*}\) induces
$$\begin{aligned} 0=-2\beta ^2\left( \beta +1\right) +5i+4=3-t \end{aligned}$$which means that \(t=3\) and \(\beta ^2=-3\beta -1\), a contradiction with the inequality (3) in Lemma 6.
As a consequence, \({\mathcal {C}}\) is an MDS \(\left( 5p,\,8\right) _{p}\) symbol-pair code. The desired result follows.
\(\square \)
Rights and permissions
About this article
Cite this article
Ma, J., Luo, J. Constructions of MDS symbol-pair codes with minimum distance seven or eight. Des. Codes Cryptogr. 90, 2337–2359 (2022). https://doi.org/10.1007/s10623-022-01081-9
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10623-022-01081-9