betaFeedback

WorksComputer Science

Linear Equivalence of Full Weight Spectrum Cyclic Orbit Codes

Abstract

We determine the linear-equivalence classes within the two families of full weight spectrum one-orbit cyclic subspace codes. This solves the second open problem posed by Castello, Polverino and Zullo [J. Combin. Theory Ser. A 212 (2025) 106005]. For the first family, the criterion is a fractional linear transformation over \F_q followed by a power of qq-Frobenius. For the second family, it is an affine transformation over \F_{q^2} together with a power of qq-Frobenius. We use polynomial spaces over finite fields, the linear-isometry theorem for cyclic orbit codes, and trace duality. The arguments include parameters belonging to proper subfields of the ambient field.

Introduction and main result

Subspace codes were introduced by Koetter and Kschischang [14] for error correction in random linear network coding. Cyclic codes in projective space were introduced by Etzion and Vardy [8], and orbit codes under cyclic subgroups of the general linear group were studied by Trautmann, Manganiello, Braun and Rosenthal [23]. Unions of cyclic orbit codes with large minimum distance were constructed by means of subspace polynomials in [3, 15, 16].

Let qq be a prime power, let n≥2n \ge2 be an integer, and put L=FqnL = \mathbb{F}_{q^n}. For an Fq\mathbb{F}_q-subspace UU of LL, write

CL(U)={aU:a∈L∗}.\mathcal{C}_L(U) = \{aU : a \in L^*\}.

This is the one-orbit cyclic subspace code, or cyclic orbit code, defined by UU. We say that CL(U)\mathcal{C}_L(U) and CL(V)\mathcal{C}_L(V) are linearly equivalent if

T(CL(U))=CL(V)for some T∈GL⁡Fq(L).T\bigl(\mathcal{C}_L(U)\bigr) = \mathcal{C}_L(V) \qquad\text{for some } T \in\operatorname{GL}_{\mathbb{F}_q}(L).

Thus the equivalence is induced by an invertible Fq\mathbb{F}_q-linear map on the ambient field. In [6, 11] linearly equivalent codes are called linearly isometric. For two kk-dimensional subspaces, the subspace distance is

d(U,V)=2k−2dim⁡Fq(U∩V).d(U,V) = 2k - 2\dim_{\mathbb{F}_q}(U \cap V).

A one-orbit cyclic subspace code has full weight spectrum (FWS) if all the distances 2,4,…,2k2, 4, \ldots, 2k occur.

Roth, Raviv and Tamo [18] observed that, for k=dim⁡FqU≥2k = \dim_{\mathbb{F}_q} U \ge2, the code CL(U)C_L(U) has (qn−1)/(q−1)(q^n - 1)/(q - 1) elements and minimum distance 2k−22k - 2 if and only if UU is a Sidon space. The equivalence of Sidon spaces was studied in [4]. Multi-orbit cyclic subspace codes were related to linear sets in [24] and used to construct optical orthogonal codes in [17].

Since d(aU,bU)=d(U,a−1bU)d(aU,bU) = d(U,a^{-1}bU) for a,b∈L∗a,b \in L^*, the distance distribution of CL(U)C_L(U) is determined by the dimensions of the intersections U∩aUU \cap aU with a∈L∗a \in L^*. Gluesing-Luerssen and Lehmann [10] studied this distribution, which is a finer invariant of cyclic orbit codes than the minimum distance. If dim⁡FqU=k\dim_{\mathbb{F}_q} U = k, the distance between two distinct codewords lies in {2,4,…,2k}\{2,4,\ldots,2k\}, and so CL(U)C_L(U) has at most kk distinct nonzero distances. The FWS codes are those attaining this bound.

The analogous question in the Hamming metric is classical; see Delsarte [7]. Shi, Zhu, Solé and Cohen [20] proved that a linear code of dimension mm over Fq\mathbb{F}_q has at most (qm−1)/(q−1)(q^m - 1)/(q - 1) distinct nonzero weights, and codes attaining this bound exist for all qq and mm by [2, 20]. A linear code of length NN has at most NN distinct nonzero weights, and Alderson [1] determined the parameters of the codes attaining this bound, which are called full weight spectrum codes. The corresponding question for cyclic codes was studied by Shi, Li, Neri and Solé [19].

Castello, Polverino and Zullo [6], Theorem 1.2 classified the FWS one-orbit cyclic subspace codes into the following two families. When k=1k = 1, the orbit is the set of all one-dimensional Fq\mathbb{F}_q-subspaces of LL, so there is only one code. We therefore assume k≥2k \ge2.

Family I. For λ∈L∖Fq\lambda\in L \setminus\mathbb{F}_q, put

Uk(λ)=Span⁡Fq{1,λ,…,λk−1},tλ=[Fq(λ):Fq].U_k(\lambda) = \operatorname{Span}_{\mathbb{F}_q}\{1,\lambda,\ldots,\lambda^{k-1}\}, \qquad t_\lambda= [\mathbb{F}_q(\lambda):\mathbb{F}_q].

The admissible range is

{2k≤n,if tλ=n,2k−1≤tλ,if tλ<n.(1)\begin{cases} 2k \le n, & \text{if } t_\lambda= n,\\ 2k - 1 \le t_\lambda, & \text{if } t_\lambda< n. \tag*{(1)} \end{cases}

Family II. Suppose nn is even and put K=Fq2⊆LK = \mathbb{F}_{q^2} \subseteq L. Let λ∈L∖K\lambda\in L \setminus K and tλ=[K(λ):K]t_\lambda= [K(\lambda):K]. For an integer rr with 1≤r≤tλ1 \le r \le t_\lambda, define

Ar(λ)=Span⁡K{1,λ,…,λr−1}.A_r(\lambda) = \operatorname{Span}_K\{1,\lambda,\ldots,\lambda^{r-1}\}.

For a positive integer ℓ\ell, put

Sℓ(λ)=Aℓ(λ)⊕Fqλℓ.S_\ell(\lambda) = A_\ell(\lambda) \oplus\mathbb{F}_q\lambda^\ell.

Here the admissible range is

tλ>2ℓ,k=2ℓ+1.(2)t_\lambda> 2\ell, \qquad k = 2\ell+ 1. \tag*{(2)}

In particular, dim⁡FqSℓ(λ)=2ℓ+1\dim_{\mathbb{F}_q} S_\ell(\lambda) = 2\ell+ 1 and [K(λ):Fq]=2tλ≥2k[K(\lambda):\mathbb{F}_q] = 2t_\lambda\ge2k.

Linear and semilinear isometries of subspace codes were studied by Trautmann [22]. For cyclic orbit codes, Gluesing-Luerssen and Lehmann [11] proved that if 1∈U1 \in U, $2 \le dim⁡FqU≤n/2\dim_{\mathbb{F}_q} U \le n/2 and UU is not contained in a proper subfield of LL, then every linear isometry from another cyclic orbit code onto CL(U)\mathcal{C}_L(U) is of the form z↦azqjz \mapsto az^{q^j} with a∈L∗a \in L^* and 0≤j<n0 \le j < n. They also observed that cyclic orbit codes with the same distance distribution need not be linearly equivalent. For k=3k=3, Castello, Polverino and Zullo [5 Theorems III.6 and III.7] determined when two spaces U3(λ)U_3(\lambda) with tλ>4t_\lambda> 4, or two spaces S1(λ)S_1(\lambda), are mapped onto each other by a map z↦azσz \mapsto az^\sigma with a∈L∗a \in L^* and σ∈Aut⁡(L)\sigma\in\operatorname{Aut}(L).

The question of determining the linear-equivalence classes of the codes in Families I and II was posed as the second open problem in the concluding remarks of [6]. In general, the classification of subspace codes up to equivalence is a difficult problem; see [13] and, for example, the classification of the optimal binary codes of length 8, constant dimension 4 and minimum distance 6 in [12]. The two families are already known to be inequivalent by [6 Corollary 4.12]. In this paper we determine the equivalence classes within each family for all k≥2k \ge2, and hence we solve this open problem. Our main result is the following.

Theorem 1.1. Fix qq, nn, kk with k≥2k \ge2.

(i) Let λ,μ∈L∖Fq\lambda, \mu\in L \setminus\mathbb{F}_q satisfy (1). The codes CL(Uk(λ))\mathcal{C}_L(U_k(\lambda)) and CL(Uk(μ))\mathcal{C}_L(U_k(\mu)) are linearly equivalent if and only if

μ=aλqj+bcλqj+d(3)\mu= \frac{a\lambda^{q^j}+b}{c\lambda^{q^j}+d} \tag*{(3)}

for some a,b,c,d∈Fqa,b,c,d \in\mathbb{F}_q with ad−bc≠0ad-bc \ne0 and some integer 0≤j<n0 \le j < n.

(ii) Suppose nn is even, K=Fq2K = \mathbb{F}_{q^2} and k=2ℓ+1k = 2\ell+ 1. Let λ,μ∈L∖K\lambda, \mu\in L \setminus K satisfy (2). The codes CL(Sℓ(λ))\mathcal{C}_L(S_\ell(\lambda)) and CL(Sℓ(μ))\mathcal{C}_L(S_\ell(\mu)) are linearly equivalent if and only if

μ=a+bλqj(4)\mu= a+b\lambda^{q^j} \tag*{(4)}

for some a∈Ka \in K, b∈K∗b \in K^* and some integer 0≤j<n0 \le j < n.

In each part, the degrees of the fields generated by λ\lambda and μ\mu over the indicated base field are then equal.

Since the two families are inequivalent and there is only one code for k=1k=1, Theorem 1.1 gives the linear-equivalence classes of all FWS one-orbit cyclic subspace codes.

Remark 1.2. Let mλ∈Fq[X]m_\lambda\in\mathbb{F}_q[X] be the minimal polynomial of λ\lambda over Fq\mathbb{F}_q. Since fractional linear transformations over Fq\mathbb{F}_q commute with the qq-Frobenius map, PGL⁡2(Fq)\operatorname{PGL}_2(\mathbb{F}_q) acts on the monic irreducible polynomials over Fq\mathbb{F}_q of a fixed degree t≥2t \ge2 through its action on their roots; see [9, 21]. Part (i) of Theorem 1.1 says that CL(Uk(λ))\mathcal{C}_L(U_k(\lambda)) and CL(Uk(μ))\mathcal{C}_L(U_k(\mu)) are linearly equivalent if and only if mμm_\mu lies in the PGL⁡2(Fq)\operatorname{PGL}_2(\mathbb{F}_q)-orbit of mλm_\lambda. In particular, the criterion does not depend on kk. If tt is a prime not dividing ∣PGL⁡2(Fq)∣=q(q2−1)|\operatorname{PGL}_2(\mathbb{F}_q)| = q(q^2 - 1), then this action is free on the polynomials of degree tt, since a nonidentity element fixing such a polynomial would act on its roots as a nontrivial power of the Frobenius map, and hence would have order tt. Hence, for each admissible kk, the codes CL(Uk(λ))C_L(U_k(\lambda)) with tλ=tt_\lambda=t form exactly (qt−1−1)/(t(q2−1))(q^{t-1}-1)/(t(q^2-1)) classes. For example, if q=2q=2 and tλ=n=7t_\lambda=n=7, then there are three classes for each k∈{2,3}k\in\{2,3\}, although for fixed kk these codes have the same distance distribution by [6 Corollary 4.6]; for k=3k=3 see also [5 Remark III.9]. Part (ii) can be stated in a similar way with minimal polynomials over KK, using the substitutions X↦a+bXX\mapsto a+bX with a∈Ka\in K, b∈K∗b\in K^* and the nontrivial automorphism of KK acting on the coefficients.

Remark 1.3. Write q=peq=p^e with pp prime. Let TT be a semilinear bijection of LL, that is, TT is additive and T(cz)=cσT(z)T(cz)=c^\sigma T(z) for all c∈Fqc\in\mathbb{F}_q and z∈Lz\in L, where σ\sigma is an automorphism of Fq\mathbb{F}_q. Then TT is the composition of an Fq\mathbb{F}_q-linear map with z↦zpiz\mapsto z^{p^i} for some ii. Since z↦zpiz\mapsto z^{p^i} maps CL(Uk(λ))C_L(U_k(\lambda)) onto CL(Uk(λpi))C_L(U_k(\lambda^{p^i})) and CL(Sℓ(λ))C_L(S_\ell(\lambda)) onto CL(Sℓ(λpi))C_L(S_\ell(\lambda^{p^i})), Theorem 1.1 also determines the equivalence of these codes under semilinear bijections if λqj\lambda^{q^j} with 0≤j<n0\leq j<n is replaced by λpj\lambda^{p^j} with 0≤j<en0\leq j<en. For k=3k=3 this agrees with [5].

The rest of this paper is organized as follows. In Section 2, we first reduce to the smallest field containing the representative subspace (Lemma 2.1). This permits us to apply Theorem 2.2, the linear-isometry theorem of Gluesing-Luerssen and Lehmann [11]. We then determine the possible scalar equivalences by polynomial calculations (Lemmas 3.2 and 4.2). For Family I, the case tλ=2k−1t_\lambda=2k-1 is treated separately using the trace pairing (Lemma 3.3). We prove part (i) of Theorem 1.1 in Section 3 and part (ii) in Section 4. Section 5 concludes the paper.

Reduction to the generated field

In this section we show that if two cyclic orbit codes are linearly equivalent, then their representatives containing 11 generate the same subfield EE of LL, and the equivalence can be restricted to EE (Lemma 2.1). Then we recall the linear-isometry theorem of Gluesing-Luerssen and Lehmann (Theorem 2.2), which we apply in EE.

For an Fq\mathbb{F}_q-subspace U⊆LU\subseteq L containing 11, let

E(U)=Fq(U)E(U)=\mathbb{F}_q(U)

be the smallest subfield of LL containing UU. The degree of this field is invariant under linear equivalence; see [6 Corollary 2.6]. The following argument also gives the restriction of an equivalence that we need.

Lemma 2.1. Let 1∈U⊆L1\in U\subseteq L and dim⁡FqU≥2\dim_{\mathbb{F}_q}U\geq2. Form a graph whose vertices are the codewords of CL(U)C_L(U), joining two distinct vertices when their intersection is nonzero. The set of distinct Fq\mathbb{F}_q-linear spans of the connected components is

{aE(U):a∈L∗}.\{aE(U):a\in L^*\}.

Consequently, if 1∈V1 \in V and CL(U)C_L(U) is linearly equivalent to CL(V)C_L(V), then E(U)=E(V)=EE(U)=E(V)=E. After multiplication by a scalar, the equivalence restricts to an Fq\mathbb{F}_q-linear equivalence between CE(U)C_E(U) and CE(V)C_E(V).

Proof. Put E=E(U)E=E(U), write U∗=U∖{0}U^*=U\setminus\{0\}, and let

H=⟨u/v:u,v∈U∗⟩H=\langle u/v:u,v\in U^*\rangle

be the multiplicative subgroup of E∗E^* generated by these quotients. For a,b∈L∗a,b\in L^*,

aU∩bU≠{0}⟺b/a∈U∗/U∗.aU\cap bU\ne\{0\}\quad\Longleftrightarrow\quad b/a\in U^*/U^*.

It follows that the component containing UU consists of the subspaces hUhU with h∈Hh\in H. Since 1∈U1\in U, we have U∗⊆HU^*\subseteq H, and hence the span of this component is

Span⁡Fq(⋃h∈HhU)=Span⁡FqH.\operatorname{Span}_{\mathbb{F}_q}\left(\bigcup_{h\in H}hU\right)=\operatorname{Span}_{\mathbb{F}_q}H.

The space on the right contains 11 and is closed under multiplication. It is a finite subring of a field, so it is a field. It contains UU and is contained in EE, and therefore it equals EE. Multiplication by any a∈L∗a\in L^* gives the assertion for the component containing aUaU. Different components may have the same span, which is why we take the set of distinct spans.

An invertible Fq\mathbb{F}_q-linear map preserves nonzero intersections, connected components and their spans. An equivalence therefore maps the set of scalar multiples of E(U)E(U) onto the corresponding set for E(V)E(V). These subspaces have the same dimension over Fq\mathbb{F}_q. Since LL has a unique subfield of each possible order, E(U)=E(V)=EE(U)=E(V)=E.

If the equivalence maps EE onto bEbE, multiply it by b−1b^{-1}. Its restriction to EE is then an invertible Fq\mathbb{F}_q-linear map of EE. Finally, the codewords of CL(U)C_L(U) contained in EE are exactly the codewords of CE(U)C_E(U): if aU⊆EaU\subseteq E, then a∈Ea\in E because 1∈U1\in U. The same holds for VV, proving the last assertion. □

We use the following form of the linear-isometry theorem [11 Theorem 6.2(a)]; see also [6 Theorem 2.3].

Theorem 2.2 (Gluesing-Luerssen–Lehmann). Let E=FqsE=\mathbb{F}_{q^s} and let rr be an integer with 2≤r≤s/22\le r\le s/2. Let U,VU,V be rr-dimensional Fq\mathbb{F}_q-subspaces of EE, with 1∈V1\in V and Fq(V)=E\mathbb{F}_q(V)=E. Then CE(U)C_E(U) and CE(V)C_E(V) are linearly equivalent if and only if

V=aUqjfor some a∈E∗ and 0≤j<s,V=aU^{q^j}\quad\text{for some }a\in E^*\text{ and }0\le j<s,

where Uqj={uqj:u∈U}U^{q^j}=\{u^{q^j}:u\in U\}.

The bound 2r≤s2r\le s is needed here. In our applications, Lemma 2.1 first replaces LL by the field generated by the subspace. We then check this dimension bound in that field.

Polynomial spaces and the first family

In this section we determine when two polynomial spaces are scalar multiples of each other (Lemma 3.2), and we prove part (i) of Theorem 1.1. For the case tλ=2k−1t_{\lambda}=2k-1 we also use the trace dual of a polynomial space given in Lemma 3.3.

Let BB be a finite field and let MM be a finite extension of BB. For x∈Mx \in M, put t=[B(x):B]t=[B(x):B]. For an integer 1≤r≤t1 \le r \le t, write

Pr(x)=Span⁡B{1,x,…,xr−1}⊆M.P_r(x)=\operatorname{Span}_B\{1,x,\ldots,x^{r-1}\}\subseteq M.

Evaluation at xx is injective on polynomials over BB of degree less than tt. The next lemma is the polynomial-intersection calculation used in [6 Lemma 4.3]. We give the proof in the range t>2r−2t>2r-2, which will also cover the boundary case below.

Lemma 3.1. Let x∈Mx \in M have degree tt over BB, and let r≥2r \ge2 with t>2r−2t>2r-2. Let f,g∈B[X]f,g \in B[X] be nonzero coprime polynomials of degrees at most r−1r-1. Put z=f(x)/g(x)z=f(x)/g(x). Then

dim⁡B(Pr(x)∩zPr(x))=r−max⁡{deg⁡f,deg⁡g}.\dim_B\left(P_r(x)\cap zP_r(x)\right)=r-\max\{\deg f,\deg g\}.

Proof. An element of the intersection has the form A(x)=zD(x)A(x)=zD(x), where A,D∈B[X]A,D \in B[X] have degrees at most $r-1. Thus

(gA−fD)(x)=0.(gA-fD)(x)=0.

Since deg⁡(gA−fD)≤2r−2<t\deg(gA-fD)\le2r-2<t, this gives the polynomial identity gA=fDgA=fD. As ff and gg are coprime, A=fhA=fh and D=ghD=gh for some h∈B[X]h \in B[X]. Conversely, every such pair gives an element of the intersection. The permitted polynomials hh have degree less than r−max⁡{deg⁡f,deg⁡g}r-\max\{\deg f,\deg g\}. Evaluation and multiplication by f(x)≠0f(x)\ne0 are injective on this space, giving the stated dimension.

Lemma 3.2. Let r≥2r \ge2 and let x,y∈Mx,y \in M have degrees at least 2r−12r-1 over BB. Then Pr(y)=αPr(x)P_r(y)=\alpha P_r(x) for some α∈M∗\alpha\in M^{*} if and only if

y=ax+bcx+dy=\frac{ax+b}{cx+d}

for some a,b,c,d∈Ba,b,c,d \in B with ad−bc≠0ad-bc\ne0.

Proof. Suppose Pr(y)=αPr(x)P_r(y)=\alpha P_r(x). Since α,αx∈Pr(y)\alpha,\alpha x \in P_r(y), we may write x=f(y)/g(y)x=f(y)/g(y) with f,g∈B[X]f,g \in B[X] nonzero and coprime, both of degree at most r−1.Thepowers$1,x,…,xrr-1. The powers \$1,x,\ldots,x^r are linearly independent over BB, so

dim⁡B(Pr(x)∩xPr(x))=r−1.\dim_B\left(P_r(x)\cap xP_r(x)\right)=r-1.

Multiplication by α\alpha and Lemma 3.1 now give

r−1=dim⁡B(Pr(y)∩xPr(y))=r−max⁡{deg⁡f,deg⁡g}.r-1=\dim_B\left(P_r(y)\cap xP_r(y)\right)=r-\max\{\deg f,\deg g\}.

Thus max⁡{deg⁡f,deg⁡g}=1\max\{\deg f,\deg g\}=1. The quotient is nonconstant because x∉Bx\notin B. Its coefficient matrix is therefore invertible, and inversion gives the required expression for yy.

Conversely, suppose y=(ax+b)/(cx+d)y=(ax+b)/(cx+d) with ad−bc≠0ad-bc\ne0. The denominator is nonzero since x∉Bx\notin B. For 0≤i<r0\le i<r,

(cx+d)r−1yi=(ax+b)i(cx+d)r−1−i∈Pr(x).(cx+d)^{r-1}y^i=(ax+b)^i(cx+d)^{r-1-i}\in P_r(x).

Both spaces have dimension rr, and hence

(cx+d)r−1Pr(y)=Pr(x).(5)(cx+d)^{r-1}P_r(y)=P_r(x). \tag*{(5)}

This proves the converse. □

For the case tλ=2k−1t_\lambda=2k-1, we shall also need the following trace-dual identity; see [6 Proposition 3.3].

Lemma 3.3. Let x∈E=Fqtx\in E=\mathbb{F}_{q^t} satisfy [Fq(x):Fq]=t[\mathbb{F}_q(x):\mathbb{F}_q]=t, and let f∈Fq[X]f\in\mathbb{F}_q[X] be its monic minimal polynomial. For 1≤k<t1\le k<t, put Uk(x)=Span⁡Fq{1,x,…,xk−1}U_k(x)=\operatorname{Span}_{\mathbb{F}_q}\{1,x,\ldots,x^{k-1}\}. With respect to the trace pairing

(u,v)⟼Tr⁡E/Fq(uv),(u,v)\longmapsto\operatorname{Tr}_{E/\mathbb{F}_q}(uv),

we have

Uk(x)⊥=f′(x)−1Ut−k(x).(6)U_k(x)^\perp=f'(x)^{-1}U_{t-k}(x). \tag*{(6)}

Here f′f' is the formal derivative of ff, and f′(x)≠0f'(x)\ne0.

Proof. The polynomial ff divides Xqt−XX^{q^t}-X, whose derivative is −1-1. Thus ff is separable and f′(x)≠0f'(x)\ne0.

For 0≤i≤t−20\le i\le t-2, Lagrange interpolation at the roots of ff gives

Xi=∑f(α)=0αif(X)(X−α)f′(α).X^i=\sum_{f(\alpha)=0}\alpha^i\frac{f(X)}{(X-\alpha)f'(\alpha)}.

Comparing coefficients of Xt−1X^{t-1} yields

Tr⁡E/Fq(xif′(x))=∑f(α)=0αif′(α)=0.\operatorname{Tr}_{E/\mathbb{F}_q}\left(\frac{x^i}{f'(x)}\right)=\sum_{f(\alpha)=0}\frac{\alpha^i}{f'(\alpha)}=0.

Products of a basis element of Uk(x)U_k(x) and a basis element of Ut−k(x)U_{t-k}(x) have the form xix^i with 0≤i≤t−20\le i\le t-2. Therefore f′(x)−1Ut−k(x)⊆Uk(x)⊥f'(x)^{-1}U_{t-k}(x)\subseteq U_k(x)^\perp. The trace pairing is nondegenerate, and both sides have dimension t−kt-k. This proves (6). □

Proof of Theorem 1.1(i). Suppose the two codes are linearly equivalent. Their representatives contain 11 and generate Fq(λ)\mathbb{F}_q(\lambda) and Fq(μ)\mathbb{F}_q(\mu), respectively. By Lemma 2.1, these fields are equal, say E=FqtE=\mathbb{F}_{q^t}, and the equivalence restricts to the orbit codes in EE.

First suppose 2k≤t2k\le t. Theorem 2.2 gives

Uk(μ)=αUk(λ)qj=αUk(λqj)U_k(\mu)=\alpha U_k(\lambda)^{q^j}=\alpha U_k(\lambda^{q^j})

for some α∈E∗\alpha\in E^* and 0≤j<t0\le j<t. Lemma 3.2, with B=FqB=\mathbb{F}_q, now gives (3).

It remains to consider t=2k−1<nt=2k-1<n. Suppose first that k≥3k\ge3. Let TT be an Fq\mathbb{F}_q-linear equivalence between CE(Uk(λ))C_E(U_k(\lambda)) and CE(Uk(μ))C_E(U_k(\mu)). Define its adjoint T∗T^* by

Tr⁡E/Fq(T(u)v)=Tr⁡E/Fq(uT∗(v)).\operatorname{Tr}_{E/\mathbb{F}_q}(T(u)v)=\operatorname{Tr}_{E/\mathbb{F}_q}(uT^*(v)).

For every subspace W⊆EW\subseteq E,

T(W)⊥=(T∗)−1(W⊥).T(W)^\perp=(T^*)^{-1}(W^\perp).

Thus (T∗)−1(T^*)^{-1} maps the dual codes onto one another. Since (aW)⊥=a−1W⊥(aW)^\perp=a^{-1}W^\perp, Lemma 3.3 identifies these dual orbit codes with

CE(Uk−1(λ))andCE(Uk−1(μ)).C_E(U_{k-1}(\lambda)) \quad\text{and}\quad C_E(U_{k-1}(\mu)).

Both representatives still contain 11 and the corresponding parameter. Moreover, 2≤k−1≤t/22\le k-1\le t/2. Applying Theorem 2.2 and Lemma 3.2 to these spaces again gives (3).

If k=2k=2 and t=3t=3, the group PGL⁡2(Fq)\operatorname{PGL}_2(\mathbb{F}_q) acts freely on the degree-three elements of Fq3\mathbb{F}_{q^3}. Indeed, a nonidentity fractional linear transformation fixing such an element would give a nonzero polynomial of degree at most two having it as a root. There are q3−qq^3-q degree-three elements, and ∣PGL⁡2(Fq)∣=q3−q|\operatorname{PGL}_2(\mathbb{F}_q)|=q^3-q. The action is therefore transitive, so any two such parameters satisfy (3) with j=0j=0.

For the converse, z↦zqjz\mapsto z^{q^j} is an invertible Fq\mathbb{F}_q-linear map on LL and sends CL(Uk(λ))C_L(U_k(\lambda)) onto CL(Uk(λqj))C_L(U_k(\lambda^{q^j})). (5) shows that the latter orbit code equals CL(Uk(μ))C_L(U_k(\mu)). This proves the assertion. □

The second family

In this section we prove part (ii) of Theorem 1.1. We show that Sℓ(λ)S_\ell(\lambda) determines the space K+KλK+K\lambda (Lemma 4.1), and using this we determine when Sℓ(μ)S_\ell(\mu) is a scalar multiple of Sℓ(λ)S_\ell(\lambda) (Lemma 4.2).

Throughout this section, nn is even, K=Fq2⊆LK=\mathbb{F}_{q^2}\subseteq L, and ℓ≥1\ell\ge1. Let λ∈L∖K\lambda\in L\setminus K with t=[K(λ):K]>2ℓt=[K(\lambda):K]>2\ell. To simplify the notation, put

A=Span⁡K{1,λ,…,λℓ−1},A=\operatorname{Span}_K\{1,\lambda,\ldots,\lambda^{\ell-1}\},
S=A⊕Fqλℓ,S=A\oplus\mathbb{F}_q\lambda^\ell,
B=Span⁡K{1,λ,…,λℓ}.B=\operatorname{Span}_K\{1,\lambda,\ldots,\lambda^\ell\}.

The next lemma shows that AA is the largest KK-linear subspace contained in SS, and that BB is the KK-linear span of SS, that is, the space of all finite sums of elements asas with a∈Ka\in K and s∈Ss\in S.

Lemma 4.1. For any ω∈K∖Fq\omega\in K\setminus\mathbb{F}_q,

S∩ωS=A,Span⁡KS=B.S\cap\omega S=A,\qquad\operatorname{Span}_K S=B.

Every KK-linear subspace of SS is contained in AA. Furthermore,

{z∈L:zA⊆B}=K+Kλ.(7)\{z\in L:zA\subseteq B\}=K+K\lambda. \tag*{(7)}

Proof. The elements 1,λ,…,λℓ1,\lambda,\ldots,\lambda^{\ell} are linearly independent over KK, and multiplication by ω\omega preserves AA. Consequently,

S∩ωS=A⊕(Fq∩ωFq)λℓ=A.S \cap\omega S = A \oplus(\mathbb{F}_q \cap\omega\mathbb{F}_q)\lambda^{\ell} = A.

If W⊆SW \subseteq S is KK-linear, then W=ωW⊆ωSW = \omega W \subseteq\omega S, and therefore W⊆AW \subseteq A. Taking the KK-span of SS gives BB.

For (7), suppose first that ℓ=1\ell= 1. Then A=KA = K and B=K+KλB = K + K\lambda, so zA⊆BzA \subseteq B is equivalent to z∈Bz \in B.

Now let ℓ≥2\ell\ge2 and suppose zA⊆BzA \subseteq B. Since 1∈A1 \in A, we have z∈Bz \in B, so z=h(λ)z = h(\lambda) for a polynomial h∈K[X]h \in K[X] of degree at most ℓ\ell. If h=0h = 0, the assertion is immediate. Otherwise, λℓ−1∈A\lambda^{\ell-1} \in A implies

h(λ)λℓ−1∈B.h(\lambda)\lambda^{\ell-1} \in B.

The polynomial h(X)Xℓ−1h(X)X^{\ell-1} has degree at most 2ℓ−1<t2\ell- 1 < t. Its value at λ\lambda belongs to BB only if this polynomial has degree at most ℓ\ell, by injectivity of evaluation below degree tt. Hence deg⁡h≤1\deg h \le1, giving z∈K+Kλz \in K + K\lambda. Conversely, each element of K+KλK + K\lambda maps AA into BB by multiplication.

For any c∈L∗c \in L^* we have

{z∈L:z(cA)⊆cB}={z∈L:zA⊆B}.(8)\{z \in L : z(cA) \subseteq cB\} = \{z \in L : zA \subseteq B\}. \tag*{(8)}

since cc can be cancelled. This gives the following scalar-equivalence criterion.

Lemma 4.2. Let λ,μ∈L∖K\lambda,\mu\in L \setminus K satisfy [K(λ):K]>2ℓ[K(\lambda):K] > 2\ell and [K(μ):K]>2ℓ[K(\mu):K] > 2\ell. Then Sℓ(μ)=cSℓ(λ)S_{\ell}(\mu) = cS_{\ell}(\lambda) for some c∈L∗c \in L^* if and only if

μ=a+bλ,a∈K,b∈K∗.\mu= a + b\lambda,\qquad a \in K,\qquad b \in K^*.

Proof. Suppose Sℓ(μ)=cSℓ(λ)S_{\ell}(\mu) = cS_{\ell}(\lambda). Multiplication by cc is KK-linear. It therefore carries the largest KK-subspace of Sℓ(λ)S_{\ell}(\lambda) onto that of Sℓ(μ)S_{\ell}(\mu), and carries their KK-spans onto one another. By Lemma 4.1,

Aℓ(μ)=cAℓ(λ),Aℓ+1(μ)=cAℓ+1(λ).A_{\ell}(\mu) = cA_{\ell}(\lambda),\qquad A_{\ell+1}(\mu) = cA_{\ell+1}(\lambda).

Equations (7) and (8) imply

K+Kμ=K+Kλ.K + K\mu= K + K\lambda.

Thus μ=a+bλ\mu= a + b\lambda for a,b∈Ka,b \in K, with b≠0b \ne0 since μ∉K\mu\notin K.

Conversely, let μ=a+bλ\mu= a + b\lambda with a∈Ka \in K and b∈K∗b \in K^*. Expanding powers gives

Aℓ(μ)=Aℓ(λ),μℓ−bℓλℓ∈Aℓ(λ).A_{\ell}(\mu) = A_{\ell}(\lambda),\qquad\mu^{\ell} - b^{\ell}\lambda^{\ell} \in A_{\ell}(\lambda).

Indeed, the change between the two polynomial bases is triangular with nonzero diagonal entries 1,b,…,bℓ−11,b,\ldots,b^{\ell-1}. Therefore

Sℓ(μ)=Aℓ(λ)⊕FqbℓLℓ=bℓSℓ(λ),S_{\ell}(\mu)=A_{\ell}(\lambda)\oplus\mathbb{F}_{q}b^{\ell}\mathcal{L}^{\ell} =b^{\ell}S_{\ell}(\lambda),

as required. □\square

Proof of Theorem 1.1(ii). The smallest field containing Sℓ(λ)S_{\ell}(\lambda) is K(λ)K(\lambda). Indeed, K⊆Sℓ(λ)K\subseteq S_{\ell}(\lambda) and λ∈Sℓ(λ)\lambda\in S_{\ell}(\lambda), also when ℓ=1\ell=1. An ambient linear equivalence therefore restricts, by Lemma 2.1, to an equivalence of the orbit codes in

E=K(λ)=K(μ)=Fq2t.E=K(\lambda)=K(\mu)=\mathbb{F}_{q^{2t}}.

Since t>2ℓt>2\ell, the codeword dimension satisfies 2ℓ+1≤t2\ell+1\leq t. Thus Theorem 2.2 applies with r=2ℓ+1r=2\ell+1 and s=2ts=2t, and gives

Sℓ(μ)=cSℓ(λ)qj=cSℓ(λqj).S_{\ell}(\mu)=cS_{\ell}(\lambda)^{q^{j}}=cS_{\ell}(\lambda^{q^{j}}).

for some c∈E∗c\in E^{*} and 0≤j<2t0\leq j<2t. The last equality holds because qq-Frobenius preserves KK and fixes Fq\mathbb{F}_{q}. Lemma 4.2 now gives (4).

Conversely, if (4) holds, then

Sℓ(μ)=bℓSℓ(λqj).S_{\ell}(\mu)=b^{\ell}S_{\ell}(\lambda^{q^{j}}).

The map z↦zqjz\mapsto z^{q^{j}} on LL sends the first orbit code onto the second. This proves the theorem. □\square

Conclusion

In this paper we determined the linear-equivalence classes of all FWS one-orbit cyclic subspace codes, and hence solved the second open problem posed by Castello, Polverino and Zullo in [6]. In Family I, two codes are linearly equivalent if and only if their parameters λ\lambda and μ\mu are related by a fractional linear transformation over Fq\mathbb{F}_{q} and a power of qq-Frobenius. In Family II, the corresponding condition is an affine transformation over Fq2\mathbb{F}_{q^{2}} and a power of qq-Frobenius.

Acknowledgements

The research of M. Shi was supported by the National Natural Science Foundation of China under Grants W2611002 and 12471490.

School of Mathematical Sciences, Anhui University, Hefei 230601, China

Email address: [email protected]

Faculty of Engineering and Natural Sciences, Sabancı University, Istanbul, Türkiye

Email address: [email protected]

References

  1. [1]T. L. Alderson, A note on full weight spectrum codes, Trans. Comb. 8 (2019), no. 3, 15–22.DOI
  2. [2]T. L. Alderson and A. Neri, Maximum weight spectrum codes, Adv. Math. Commun. 13 (2019), no. 1, 101–119.
  3. [3]E. Ben-Sasson, T. Etzion, A. Gabizon and N. Raviv, Subspace polynomials and cyclic subspace codes, IEEE Trans. Inform. Theory 62 (2016), no. 3, 1157–1165.
  4. [4]C. Castello, O. Polverino, P. Santonastaso and F. Zullo, Constructions and equivalence of Sidon spaces, J. Algebraic Combin. 58 (2023), no. 4, 1299–1329.arxiv.org/abs/2303.17306
  5. [5]C. Castello, O. Polverino and F. Zullo, On one-orbit cyclic subspace codes of G_q(n, 3), in: Proc. 2024 IEEE Int. Symp. Inf. Theory (ISIT), Athens, 2024, pp. 2383–2388.
  6. [6]C. Castello, O. Polverino and F. Zullo, Full weight spectrum one-orbit cyclic subspace codes, J. Combin. Theory Ser. A 212 (2025), 106005.arxiv.org/abs/2405.19911
  7. [7]Ph. Delsarte, Four fundamental parameters of a code and their combinatorial significance, Information and Control 23 (1973), no. 5, 407–438.DOI
  8. [8]T. Etzion and A. Vardy, Error-correcting codes in projective space, IEEE Trans. Inform. Theory 57 (2011), no. 2, 1165–1173.
  9. [9]T. Garefalakis, On the action of GL_2(F_q) on irreducible polynomials over F_q, J. Pure Appl. Algebra 215 (2011), no. 8, 1835–1843.
  10. [10]H. Gluesing-Luerssen and H. Lehmann, Distance distributions of cyclic orbit codes, Des. Codes Cryptogr. 89 (2021), no. 3, 447–470.
  11. [11]H. Gluesing-Luerssen and H. Lehmann, Automorphism groups and isometries for cyclic orbit codes, Adv. Math. Commun. 17 (2023), no. 1, 119–138.
  12. [12]D. Heinlein, T. Honold, M. Kiermaier, S. Kurz and A. Wassermann, Classifying optimal binary subspace codes of length 8, constant dimension 4 and minimum distance 6, Des. Codes Cryptogr. 87 (2019), no. 2–3, 375–391.
  13. [13]T. Honold, M. Kiermaier and S. Kurz, Constructions and bounds for mixed-dimension subspace codes, Adv. Math. Commun. 10 (2016), no. 3, 649–682.arxiv.org/abs/1512.06660
  14. [14]R. Koetter and F. R. Kschischang, Coding for errors and erasures in random network coding, IEEE Trans. Inform. Theory 54 (2008), no. 8, 3579–3591.
  15. [15]K. Otal and F. Özbudak, Cyclic subspace codes via subspace polynomials, Des. Codes Cryptogr. 85 (2017), no. 2, 191–204.DOI
  16. [16]K. Otal and F. Özbudak, Constructions of cyclic subspace codes and maximum rank distance codes, in: Network Coding and Subspace Designs, Signals Commun. Technol., Springer, Cham, 2018, pp. 43–66.DOI
  17. [17]F. Özbudak, P. Santonastaso and F. Zullo, Using multi-orbit cyclic subspace codes for constructing optical orthogonal codes, Cryptogr. Commun. 17 (2025), no. 4, 695–708.arxiv.org/abs/2405.18996
  18. [18]R. M. Roth, N. Raviv and I. Tamo, Construction of Sidon spaces with applications to coding, IEEE Trans. Inform. Theory 64 (2018), no. 6, 4412–4422.arxiv.org/abs/1705.04560
  19. [19]M. Shi, X. Li, A. Neri and P. Solé, How many weights can a cyclic code have?, IEEE Trans. Inform. Theory 66 (2020), no. 3, 1449–1459.arxiv.org/abs/1807.08418
  20. [20]M. Shi, H. Zhu, P. Solé and G. D. Cohen, How many weights can a linear code have?, Des. Codes Cryptogr. 87 (2019), no. 1, 87–95.DOI
  21. [21]H. Stichtenoth and A. Topuzoglu, Factorization of a class of polynomials over finite fields, Finite Fields Appl. 18 (2012), no. 1, 108–122.DOI
  22. [22]A.-L. Trautmann, Isometry and automorphisms of constant dimension codes, Adv. Math. Commun. 7 (2013), no. 2, 147–160.arxiv.org/abs/1205.5465
  23. [23]A.-L. Trautmann, F. Manganiello, M. Braun and J. Rosenthal, Cyclic orbit codes, IEEE Trans. Inform. Theory 59 (2013), no. 11, 7386–7404.DOI
  24. [24]F. Zullo, Multi-orbit cyclic subspace codes and linear sets, Finite Fields Appl. 87 (2023), 102153.arxiv.org/abs/2112.08781

Paper details

vlnc:W_3fnph2dj
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC0-1.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents