Papers
Topics
Authors
Recent
Search
2000 character limit reached

Twisted GRS (TGRS) Codes

Updated 18 March 2026
  • Twisted GRS (TGRS) Codes are a family of MDS and NMDS linear codes constructed by twisting the message polynomial space of generalized Reed–Solomon codes to yield non-GRS structures.
  • They feature explicit generator matrices and decoding algorithms using algebraic error-correcting pairs, Gaussian elimination, and Schur product analysis.
  • Their rich duality, self-dual constructions, and complete weight distribution properties offer practical insights for error correction and cryptographic applications.

Twisted GRS (TGRS) Codes are a family of maximum distance separable (MDS) and near-MDS (NMDS) linear codes over finite fields, constructed as extensions of classical Generalized Reed–Solomon (GRS) codes by algebraically "twisting" the message polynomial space. These codes play a central role in algebraic coding theory, yielding large sets of non-GRS MDS codes with explicit algebraic structure, families of Hermitian self-dual codes, near-MDS codes, complementary dual (LCD) codes, and possessing rich duality, error-locator, decoding, and combinatorial properties (Li et al., 4 Aug 2025, Zhu et al., 2022).

1. Formalism, Construction, and Generator Matrix

A TGRS code is defined by selecting a collection of distinct field elements a=(a1,,an)Fq\mathbf{a}=(a_1,\dots,a_n)\subset F_q, a vector of nonzero multipliers v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n, and a "twisted" polynomial message space. For the (+)(+)–TGRS (“plus-twisted”) case (the archetypal single-twist construction), the polynomial space is

Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},

i.e. fkf_k is algebraically tied to fk1f_{k-1} by the twist parameter η\eta. The TGRS code is the evaluation code

TGRSk(a,v,η)={(v1f(a1),...,vnf(an)):fVq[x]k}Fqn,\mathrm{TGRS}_k(\mathbf{a}, \mathbf{v}, \eta) = \{ (v_1f(a_1), ..., v_n f(a_n)) : f \in \mathcal{V}_q[x]_k \} \subset F_q^n,

with generator matrix GG in standard “twisted” form: the first k1k-1 rows are as in GRS, and the v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n0-th is v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n1 (Li et al., 4 Aug 2025, Zhu et al., 2022). The v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n2–extended TGRS (ETGRS) code appends the coefficient v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n3, yielding an v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n4 code.

More general “multi-twist” constructions and arbitrary-twist variants (A-TGRS, v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n5–TGRS) are parameterized by discrete twist sets, hook sets, and matrices, generalizing the single-twist case and subsuming all known TGRS families (Zhao et al., 2024, Hu et al., 7 Feb 2025).

2. MDS and NMDS Property, Weight Distribution

A v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n6–TGRS v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n7 code is MDS (i.e., minimum Hamming distance v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n8) if and only if

v=(v1,,vn)(Fq)n\mathbf{v}=(v_1,\dots,v_n)\in (F_q^*)^n9

Otherwise, the code is NMDS with (+)(+)0.

For the extended TGRS, the analogous condition is

(+)(+)1

for MDS (Zhu et al., 2022, Li et al., 4 Aug 2025). The entire weight distribution is computed in closed form based on the (+)(+)2-subset sum structure (Li et al., 4 Aug 2025).

3. Non-GRS Structure and Schur Product

Almost all TGRS/ETGRS codes, when MDS, are not equivalent to any GRS code over (+)(+)3, as established by explicit Schur square (coordinatewise product) dimension arguments. For (+)(+)4,

(+)(+)5

Dually, the Schur square of the dual has distance 1 for high rates. This property holds for all (+)(+)6 or (+)(+)7 in TGRS, and for all (+)(+)8 in ETGRS (Li et al., 4 Aug 2025). The (+)(+)9–ETGRS code is not GRS or EGRS for Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},0 (Zhu et al., 2022).

4. Duality, Self-Dual, and Self-Orthogonal Codes

Euclidean and Hermitian duals of TGRS and ETGRS admit explicit matrix expressions. The dual of ETGRS has a parity-check matrix composed of weighted evaluations and twist-dependent terms (Li et al., 4 Aug 2025, Zhu et al., 2022). Hermitian self-dual TGRS and non-GRS MDS Hermitian self-dual TGRS codes are constructed by special choice of Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},1 satisfying Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},2 over Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},3, yielding two major explicit classes (Li et al., 4 Aug 2025). There are no Galois self-dual ETGRS codes of length Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},4 and dimension Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},5.

For Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},6–TGRS with even Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},7 and suitable parameters, self-dual and almost self-dual families exist by choice of evaluation set and twisting values. Explicit criteria (based on evaluation point sum) and construction algorithms guarantee self-duality and prescribe multipliers (Zhu et al., 2022).

5. Decoding Algorithms: Error-Correcting Pairs, Gaussian Elimination, List Decoding

Decoding TGRS and ETGRS codes employs algebraic error-correcting pairs (ECPs) and enables fast decoding. ETGRS codes always admit suitable ECPs: if Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},8 is odd, an Vq[x]k={f(x)=i=0k1fixi+ηfk1xk:fiFq,ηFq},\mathcal{V}_q[x]_k = \left\{f(x) = \sum_{i=0}^{k-1} f_i x^i + \eta f_{k-1} x^k : f_i \in F_q,\, \eta \in F_q^* \right\},9–ECP exists; if even, an fkf_k0–ECP. This allows polynomial-time (often fkf_k1) decoders correcting up to half the minimum distance. MDS fkf_k2–TGRS codes with fkf_k3 do not admit fkf_k4–ECPs, distinguishing them from GRS (Li et al., 4 Aug 2025). Decoders based on Gaussian elimination solve certain structured polynomial systems associated with received words for both MDS and NMDS TGRS codes (Zhang et al., 5 Aug 2025). Complexity compares favorably to classical polynomial-time GRS decoders, and handles arbitrary twist positions.

There is no unique decoding algorithm based on ECPs for MDS fkf_k5–TGRS codes when certain symmetry conditions are met, as any such would incorrectly imply GRS equivalence (Li et al., 4 Aug 2025).

6. Covering Radius, Deep Holes, and Extension Theory

For MDS TGRS codes, the covering radius of the dual code is exactly fkf_k6 (the code dimension), and explicit fkf_k7-affine families of deep holes are described. The fkf_k8–ETGRS code can be interpreted as a second extension of fkf_k9–TGRS and remains MDS if and only if a vector appended is a deep hole. The duals of TGRS codes constructed by Han and Zhang are shown to have covering radius fk1f_{k-1}0 and a complete family of deep holes via affine shifts (Li et al., 4 Aug 2025). This completes the analysis of maximal-likelihood decoding and geometric structure for non-GRS MDS codes in the TGRS/ETGRS family.

7. Concrete Examples and Practical Implications

Explicit small-field examples demonstrate all phenomena, including MDS/NMDS status, non-GRS nature, construction of ECPs and execution of decoding algorithms, and enumeration of deep holes. For instance, for fk1f_{k-1}1, fk1f_{k-1}2, fk1f_{k-1}3, fk1f_{k-1}4, fk1f_{k-1}5, fk1f_{k-1}6, fk1f_{k-1}7, so fk1f_{k-1}8 is fk1f_{k-1}9–MDS and not GRS by Schur-square computation (Li et al., 4 Aug 2025).

Table: Key Theoretical Properties of η\eta0–TGRS and ETGRS Codes

Property Conditions Reference Section
MDS η\eta1 [(Li et al., 4 Aug 2025), §3]
Non-GRS MDS η\eta2 or η\eta3 [(Li et al., 4 Aug 2025), §4]
Hermitian self-dual MDS η\eta4 (over η\eta5), explicit sets [(Li et al., 4 Aug 2025), §5]
Dual code structure Explicit parity-check matrix incl. twist terms [(Li et al., 4 Aug 2025), §6]
ECP-based decoding ECP exists for ETGRS for all η\eta6 [(Li et al., 4 Aug 2025), §7]
Covering radius η\eta7 [(Li et al., 4 Aug 2025), §8]

References

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Twisted GRS (TGRS) Codes.