A note on explicit homological invariants of graded double Ore extensions Andrés Rubiano Abstract . We compute explicit homological invariants of a trimmed graded double Ore extension of the quantum plane. For a pilot family of type ( 14641 ) (14641) , we determine the minimal graded free resolution and graded Betti numbers of the trivial right module and also compute linear resolutions for two natural cyclic right modules. This provides a concrete link between the PBW structure of the algebra and the homological behavior of its natural quotients. 2020 Mathematics Subject Classification 16S36, 16E05, 16E65, 16S38 1. Introduction Artin-Schelter regular algebras, introduced in [ 1 ] , are among the basic graded analogues of polynomial rings in noncommutative algebra and projective geometry; see also [ 2 ] . A particularly effective construction of higher-dimensional regular algebras is the double Ore extension of Zhang and Zhang [ 10 ] , who proved that a connected graded double Ore extension of an Artin-Schelter regular algebra is again Artin-Schelter regular. The four-dimensional case is especially significant. In [ 11 ] , Zhang and Zhang used double Ore extensions to construct regular algebras of type (14641) , proving that they are strongly noetherian, Auslander regular, Koszul and Cohen-Macaulay. Many of these algebras are genuinely new, in the sense that they are not isomorphic to an Ore extension or a normal extension of a three-dimensional regular algebra. This makes trimmed graded double Ore extensions a natural testing ground for explicit homological questions. The existence of a finite linear resolution of the trivial module is guaranteed abstractly in the Koszul regular setting, but writing down the differentials explicitly is far subtler; see [ 9 , 8 ] . Several results point toward such an explicit program. Zhu, Van Oystaeyen and Zhang studied trimmed double Ore extensions of Koszul regular algebras [ 12 ] ; Gómez and Suárez related graded double Ore extensions to graded skew PBW extensions [ 5 ] ; and Carvalho, Lopes and Matczuk characterized when a double Ore extension collapses to an iterated Ore extension [ 3 ] . More recently, Cao, Shen and Wang developed a general homological framework for graded double Ore extensions of Koszul algebras [ 4 ] . In this paper we carry out a fully explicit computation for a concrete pilot family. We study the family K in the classification of [ 11 ] , specialized at q=-1 , over the quantum plane. Recent work of Herrera, Higuera and the author shows that this family admits a finite Gröbner-Shirshov basis and hence a PBW basis [ 6 ] . Our main result is an explicit minimal graded free resolution of the trivial right module k_{B} , written in matrix form. We also compute the minimal linear resolutions of the natural cyclic right modules M_{x}:=B/(x_{1},x_{2})B,\ M_{y}:=B/(y_{1},y_{2})B, and deduce their graded Betti numbers and Poincaré series. These computations show that, even when the total Betti pattern of the trivial module is forced by regularity, the actual differentials and the behavior of natural cyclic quotients still retain concrete information about the defining double Ore data. Throughout the paper, \Bbbk denotes a field. All algebras are associative \Bbbk -algebras and all graded algebras are connected \mathbb{N} -graded and generated in degree one unless otherwise stated. 2. Preliminaries 2.1. Double Ore extensions We recall the notion of double Ore extension introduced by Zhang and Zhang [ 10 ] . Since the original compatibility conditions contain minor misprints, we follow the corrected formulation given by Carvalho, Lopes and Matczuk [ 3 ] . Definition 2.1 ( [ 10 , Definition 1.3] ; [ 3 , Definition 1.1] ) . Let R be a subalgebra of a \Bbbk -algebra B . (a) We say that B is a right double extension of R if: (i) B is generated by R together with two new indeterminates y_{1},y_{2} ; (ii) the generators y_{1},y_{2} satisfy a relation y_{2}y_{1}=p_{12}y_{1}y_{2}+p_{11}y_{1}^{2}+\tau_{1}y_{1}+\tau_{2}y_{2}+\tau_{0}, (2.1) for some p_{12},p_{11}\in\Bbbk and \tau_{0},\tau_{1},\tau_{2}\in R ; (iii) B is a free left R -module with basis \{y_{1}^{i}y_{2}^{j}\mid i,j\geq 0\} ; (iv) y_{1}R+y_{2}R+R\subseteq Ry_{1}+Ry_{2}+R . (b) A right double extension B of R is called a double extension of R if, in addition, (i) p_{12}\neq 0 ; (ii) B is a free right R -module with basis \{y_{2}^{i}y_{1}^{j}\mid i,j\geq 0\} ; (iii) y_{1}R+y_{2}R+R=Ry_{1}+Ry_{2}+R . Condition 2.1 (a)(iv) is equivalent to the existence of maps \sigma=\begin{bmatrix}\sigma_{11}&\sigma_{12}\\ \sigma_{21}&\sigma_{22}\end{bmatrix}:R\to M_{2}(R),\ \delta=\begin{bmatrix}\delta_{1}\\ \delta_{2}\end{bmatrix}:R\to M_{2\times 1}(R), such that \begin{bmatrix}y_{1}\\ y_{2}\end{bmatrix}r=\sigma(r)\begin{bmatrix}y_{1}\\ y_{2}\end{bmatrix}+\delta(r)\ \text{for all }r\in R. (2.2) If B is a right double extension, we write B=R_{P}[y_{1},y_{2};\sigma,\delta,\tau], where P=(p_{12},p_{11})\in\Bbbk^{2} and \tau=\{\tau_{0},\tau_{1},\tau_{2}\}\subseteq R . The collection \{P,\sigma,\delta,\tau\} is called the DE-data of B . A particularly important subclass is that of trimmed double extensions , where \delta=0 and \tau=\{0,0,0\} ; then we simply write R_{P}[y_{1},y_{2};\sigma] . For a right double extension R_{P}[y_{1},y_{2};\sigma,\delta,\tau] , the map \sigma is an algebra homomorphism and \delta is a \sigma -derivation; see [ 10 , Lemma 1.7] . The corresponding compatibility conditions are given in [ 3 , Proposition 1.5] . The following criterion will be used repeatedly to separate genuine double Ore extensions from iterated Ore extensions. Proposition 2.2 ( [ 3 , Theorems 2.2 and 2.4] ) . Let B=R_{P}[y_{1},y_{2};\sigma,\delta,\tau] be a right double extension of R . (1) The following are equivalent: (i) B can be presented as an iterated Ore extension R[y_{1};\sigma_{1},d_{1}][y_{2};\sigma_{2},d_{2}] ; (ii) \sigma_{12}=0 . Under these equivalent conditions, B is a double extension if and only if p_{12}\neq 0 and \sigma_{11},\sigma_{22} are automorphisms of R . (2) The algebra B admits an iterated Ore presentation R[y_{2};\sigma_{2}^{\prime},d_{2}^{\prime}][y_{1};\sigma_{1}^{\prime},d_{1}^{\prime}] if and only if \sigma_{21}=0 , p_{12}\neq 0 and p_{11}=0 . In that situation, B is a double extension if and only if the induced maps \sigma_{22} and \sigma_{11} are automorphisms of R . 2.2. The graded setting Our main interest lies in trimmed graded double Ore extensions over an Artin-Schelter regular algebra of global dimension 2 . We therefore write A=\Bbbk_{Q}[x_{1},x_{2}]=\Bbbk\langle x_{1},x_{2}\rangle\Big/\langle x_{2}x_{1}-q_{12}x_{1}x_{2}-q_{11}x_{1}^{2}\rangle, where Q=(q_{12},q_{11})\in\Bbbk^{2} with q_{12}\neq 0 . The most relevant special cases are the quantum plane, corresponding to Q=(q,0) and the Jordan plane, corresponding to Q=(1,1) . In the connected graded Artin-Schelter regular setting of global dimension 2 , these are precisely the possibilities; see [ 11 , Lemma 2.4] . Let B=A_{P}[y_{1},y_{2};\sigma] be a trimmed graded double extension, where P=(p_{12},p_{11})\in\Bbbk^{2} with p_{12}\neq 0 . Since \sigma is graded, each \sigma_{ij}(x_{s}) is a linear combination of x_{1},x_{2} , say \sigma_{ij}(x_{s})=\sum_{t=1}^{2}a_{ijst}x_{t},\ a_{ijst}\in\Bbbk. (2.3) Thus B is defined by the two non-mixing quadratic relations \displaystyle x_{2}x_{1} \displaystyle=q_{12}x_{1}x_{2}+q_{11}x_{1}^{2}, (2.4) \displaystyle y_{2}y_{1} \displaystyle=p_{12}y_{1}y_{2}+p_{11}y_{1}^{2}, (2.5) together with the four mixing relations y_{i}x_{s}=\sigma_{i1}(x_{s})y_{1}+\sigma_{i2}(x_{s})y_{2},\ i,s\in\{1,2\}. (2.6) Zhang and Zhang showed that double extensions over a two-dimensional regular base naturally produce four-dimensional regular algebras of type (14641) . Proposition 2.3 ( [ 11 , Theorem 0.1] ) . Let B be a connected graded algebra generated by four degree-one elements. If B=A_{P}[y_{1},y_{2};\sigma] is a double extension of an Artin-Schelter regular algebra A of global dimension 2 , then: (1) B is strongly noetherian, Auslander regular, Cohen-Macaulay and a domain; (2) B is Artin-Schelter regular of global dimension 4 and of type (14641) ; in particular, B is Koszul; (3) if B is not isomorphic to an Ore extension of an Artin-Schelter regular algebra of global dimension 3 , then the corresponding trimmed double extension belongs to one of the 26 families classified in [ 11 ] . 2.3. Minimal graded free resolutions and Betti numbers From this point on, unless explicitly stated otherwise, all modules are finitely generated graded right modules. If B=\bigoplus_{n\geq 0}B_{n} is a connected graded algebra, we denote by B_{\geq 1}=\bigoplus_{n\geq 1}B_{n} its irrelevant ideal and by \varepsilon:B\to\Bbbk the canonical augmentation. The corresponding trivial right module is denoted by k_{B} . Let M be a graded right B -module. A graded free resolution of M is an exact complex \cdots\longrightarrow F_{i}\xrightarrow{d_{i}}F_{i-1}\longrightarrow\cdots\longrightarrow F_{1}\xrightarrow{d_{1}}F_{0}\longrightarrow M\longrightarrow 0, where each F_{i} is a graded free right B -module. Such a resolution is minimal if d_{i}(F_{i})\subseteq F_{i-1}B_{\geq 1}\ \text{for all }i\geq 1. In that case, each F_{i} decomposes uniquely up to graded isomorphism as F_{i}\cong\bigoplus_{j\in\mathbb{Z}}B(-j)^{\beta_{i,j}^{B}(M)}, where B(-j) denotes the degree shift. The integers \beta_{i,j}^{B}(M):=\dim_{\Bbbk}\operatorname{Tor}_{i}^{B}(M,\Bbbk)_{j} are called the graded Betti numbers of M and \beta_{i}^{B}(M):=\sum_{j}\beta_{i,j}^{B}(M) are the total Betti numbers . Definition 2.4 ( [ 7 , Chs. 11–12] ) . Let B be a connected \mathbb{N} -graded algebra and let M be a finitely generated graded right B -module. Suppose that \cdots\longrightarrow F_{i}\longrightarrow F_{i-1}\longrightarrow\cdots\longrightarrow F_{1}\longrightarrow F_{0}\longrightarrow M\longrightarrow 0 is the minimal graded free resolution of M , where F_{i}\cong\bigoplus_{j\in\mathbb{Z}}B(-j)^{\beta_{i,j}^{B}(M)}. The bigraded Poincaré series of M over B is P_{M}^{B}(s,t):=\sum_{i\geq 0}\sum_{j\in\mathbb{Z}}\beta_{i,j}^{B}(M)\,s^{i}t^{j}. Equivalently, P_{M}^{B}(s,t) is the generating series of the graded Betti numbers of M . Remark 2.5 . If one forgets the internal grading, then the ordinary Poincaré series is recovered by P_{M}^{B}(z)=\sum_{i\geq 0}\beta_{i}^{B}(M)\,z^{i},\ \text{where}\ \beta_{i}^{B}(M)=\sum_{j}\beta_{i,j}^{B}(M). In particular, the bigraded Poincaré series refines the usual Poincaré series by recording both homological degree and internal degree. If B is a Koszul Artin-Schelter regular algebra of type (14641) , then the minimal graded free resolution of k_{B} has the form 0\longrightarrow B(-4)\longrightarrow B(-3)^{4}\longrightarrow B(-2)^{6}\longrightarrow B(-1)^{4}\longrightarrow B\longrightarrow k_{B}\longrightarrow 0. (2.7) In particular, the total Betti numbers of k_{B} are (1,4,6,4,1). Therefore, the real homological interest of the present setting lies not merely in recovering the expected Betti pattern of the trivial module, but in making the differentials explicit and in comparing them with those of natural cyclic modules. For later use, recall also the standard quadratic-Koszul setup; see [ 8 , Chs. 2 and 3] . If B=T(V)/(R) is a quadratic algebra, then the right Koszul complex is built from the subspaces W_{0}=\Bbbk,\ W_{1}=V,\ W_{2}=R, and, for n\geq 3 , W_{n}=\bigcap_{i+j+2=n}V^{\otimes i}\otimes R\otimes V^{\otimes j}\subseteq V^{\otimes n}. When B is Koszul, the minimal graded free resolution of k_{B} is precisely the right Koszul complex \cdots\longrightarrow W_{n}\otimes B(-n)\longrightarrow W_{n-1}\otimes B(-(n-1))\longrightarrow\cdots\longrightarrow W_{1}\otimes B(-1)\longrightarrow B\longrightarrow k_{B}\longrightarrow 0. 3. The pilot family and its PBW structure We now fix the concrete family on which the paper is based. 3.1. The family K over the quantum plane Following [ 11 , Section 3] , we choose the double extension of type K , specialized at q=-1 . Let A=\Bbbk_{-1}[x_{1},x_{2}]=\Bbbk\langle x_{1},x_{2}\rangle/\langle x_{2}x_{1}+x_{1}x_{2}\rangle. Fix \alpha\in\Bbbk^{\times} and define B:=B_{K}(\alpha):=\Bbbk\langle x_{1},x_{2},y_{1},y_{2}\rangle/I_{K}(\alpha), where I_{K}(\alpha) is generated by \displaystyle x_{2}x_{1} \displaystyle=-x_{1}x_{2}, (3.1) \displaystyle y_{2}y_{1} \displaystyle=-y_{1}y_{2}, (3.2) \displaystyle y_{1}x_{1} \displaystyle=x_{1}y_{1}, (3.3) \displaystyle y_{1}x_{2} \displaystyle=x_{2}y_{2}, (3.4) \displaystyle y_{2}x_{1} \displaystyle=x_{1}y_{2}, (3.5) \displaystyle y_{2}x_{2} \displaystyle=\alpha x_{2}y_{1}. (3.6) These are precisely the defining relations of family K in the q=-1 case; see [ 6 , Section 4.6] . The relations ( 3.3 )-( 3.6 ) determine a graded algebra homomorphism \sigma=\begin{bmatrix}\sigma_{11}&\sigma_{12}\\ \sigma_{21}&\sigma_{22}\end{bmatrix}:A\to M_{2}(A) by \displaystyle\sigma_{11}(x_{1}) \displaystyle=x_{1}, \displaystyle\sigma_{12}(x_{1}) \displaystyle=0, \displaystyle\sigma_{21}(x_{1}) \displaystyle=0, \displaystyle\sigma_{22}(x_{1}) \displaystyle=x_{1}, \displaystyle\sigma_{11}(x_{2}) \displaystyle=0, \displaystyle\sigma_{12}(x_{2}) \displaystyle=x_{2}, \displaystyle\sigma_{21}(x_{2}) \displaystyle=\alpha x_{2}, \displaystyle\sigma_{22}(x_{2}) \displaystyle=0. Hence B=A_{P}[y_{1},y_{2};\sigma],\hskip 18.49988ptP=(-1,0), so B is a trimmed graded double Ore extension of the quantum plane. Equivalently, the corresponding matrix \Sigma_{K}(\alpha) is \Sigma_{K}(\alpha)=\begin{bmatrix}1&0&0&0\\ 0&0&0&1\\ 0&0&1&0\\ 0&\alpha&0&0\end{bmatrix}, and therefore \det(\Sigma_{K}(\alpha))=-\alpha\neq 0 . Proposition 3.1 . Let B=B_{K}(\alpha) with \alpha\neq 0 . Then: (1) B is a trimmed graded double Ore extension of the quantum plane A=\Bbbk_{-1}[x_{1},x_{2}] ; (2) B is not an iterated Ore extension in either of the two natural orders; (3) B is a Koszul Artin-Schelter regular algebra of type (14641) . Proof. The first assertion follows directly from the preceding discussion. For the second, note that \sigma_{12}(x_{2})=x_{2}\neq 0\hskip 18.49988pt\text{and}\hskip 18.49988pt\sigma_{21}(x_{2})=\alpha x_{2}\neq 0. Hence \sigma_{12}\neq 0 and \sigma_{21}\neq 0 . By Proposition 2.2 , an iterated Ore presentation in the order A[y_{1}][y_{2}] would force \sigma_{12}=0 , whereas an iterated Ore presentation in the order A[y_{2}][y_{1}] would force \sigma_{21}=0 and p_{11}=0 . Therefore neither presentation is possible. The third assertion follows from Proposition 2.3 . ∎ 3.2. PBW basis and normal monomials The family K admits a finite Gröbner-Shirshov basis and hence a PBW basis [ 6 , Section 4.6] . In the present notation, this yields the following concrete normal form. Proposition 3.2 . The algebra B has a PBW basis consisting of the ordered monomials \mathcal{B}:=\{x_{1}^{a}x_{2}^{b}y_{1}^{c}y_{2}^{d}\mid a,b,c,d\in\mathbb{N}\}. In particular, every element of B can be written uniquely as a finite \Bbbk -linear combination of monomials in \mathcal{B} and H_{B}(t)=\frac{1}{(1-t)^{4}}. Proof. By [ 6 , Section 4.6] , the defining set associated with family K is a finite Gröbner-Shirshov basis. Therefore the irreducible words form a \Bbbk -basis of B and those irreducible words are precisely the ordered monomials x_{1}^{a}x_{2}^{b}y_{1}^{c}y_{2}^{d} . The formula for the Hilbert series follows immediately. ∎ The PBW basis also makes it easy to identify the first natural cyclic quotients attached to the presentation of B . Set M_{x}:=B/(x_{1},x_{2})B,\ M_{y}:=B/(y_{1},y_{2})B. Since the mixed relations show that (x_{1},x_{2})B and (y_{1},y_{2})B are in fact two-sided ideals, these quotients may also be regarded as graded algebra quotients. In particular, one has M_{x}\cong\Bbbk_{-1}[y_{1},y_{2}],\ M_{y}\cong\Bbbk_{-1}[x_{1},x_{2}], and hence H_{M_{x}}(t)=H_{M_{y}}(t)=\frac{1}{(1-t)^{2}}. 4. Explicit minimal resolutions We now pass to the main homological computations. Throughout this section, all modules are graded right B -modules. 4.1. The trivial right module Since B is a quadratic Koszul algebra of type (14641) , the minimal graded free resolution of k_{B} has the shape 0\longrightarrow B(-4)\xrightarrow{d_{4}}B(-3)^{4}\xrightarrow{d_{3}}B(-2)^{6}\xrightarrow{d_{2}}B(-1)^{4}\xrightarrow{d_{1}}B\longrightarrow k_{B}\longrightarrow 0. Let F_{1}=B(-1)^{4} with basis e_{1},e_{2},e_{3},e_{4} , corresponding respectively to x_{1},x_{2},y_{1},y_{2} . Then d_{1}=\begin{pmatrix}x_{1}&x_{2}&y_{1}&y_{2}\end{pmatrix}. (4.1) Let F_{2}=B(-2)^{6} with basis f_{1},\dots,f_{6} corresponding to the six quadratic relations \displaystyle r_{1} \displaystyle:=x_{2}x_{1}+x_{1}x_{2}, \displaystyle r_{2} \displaystyle:=y_{2}y_{1}+y_{1}y_{2}, \displaystyle r_{3} \displaystyle:=y_{1}x_{1}-x_{1}y_{1}, \displaystyle r_{4} \displaystyle:=y_{1}x_{2}-x_{2}y_{2}, \displaystyle r_{5} \displaystyle:=y_{2}x_{1}-x_{1}y_{2}, \displaystyle r_{6} \displaystyle:=y_{2}x_{2}-\alpha x_{2}y_{1}. Then the second differential is d_{2}=\begin{pmatrix}x_{2}&0&-y_{1}&0&-y_{2}&0\\ x_{1}&0&0&-y_{2}&0&-\alpha y_{1}\\ 0&y_{2}&x_{1}&x_{2}&0&0\\ 0&y_{1}&0&0&x_{1}&x_{2}\end{pmatrix}. (4.2) By construction, the j -th column of d_{2} is the coefficient vector of r_{j} with respect to the ordered basis (x_{1},x_{2},y_{1},y_{2}) and therefore d_{1}d_{2}=0 . The next step is controlled by the cubic overlaps of the defining relations. In the monic Gröbner-Shirshov presentation of family K , there are exactly four critical cubic overlaps, all of them trivial modulo the defining relations; see [ 6 , Section 4.6] . This suggests the following four cubic syzygies: \displaystyle\eta_{1} \displaystyle:=f_{1}y_{2}+f_{3}x_{2}+f_{4}x_{1}, (4.3) \displaystyle\eta_{2} \displaystyle:=\alpha f_{1}y_{1}+f_{5}x_{2}+f_{6}x_{1}, (4.4) \displaystyle\eta_{3} \displaystyle:=f_{2}x_{1}-f_{3}y_{2}-f_{5}y_{1}, (4.5) \displaystyle\eta_{4} \displaystyle:=f_{2}x_{2}-\alpha f_{4}y_{1}-f_{6}y_{2}. (4.6) A direct verification using ( 3.1 )-( 3.6 ) shows that d_{2}(\eta_{i})=0 for i=1,2,3,4 . For example, \displaystyle d_{2}(\eta_{1}) \displaystyle=(x_{2}x_{1}+x_{1}x_{2})y_{2}+(y_{1}x_{1}-x_{1}y_{1})x_{2}+(y_{1}x_{2}-x_{2}y_{2})x_{1} \displaystyle\equiv-x_{1}x_{2}y_{2}+x_{1}x_{2}y_{2}+x_{1}y_{1}x_{2}-x_{1}y_{1}x_{2}-x_{2}y_{2}x_{1}+x_{2}y_{2}x_{1}=0. The other three cases are analogous. Let F_{3}=B(-3)^{4} with basis g_{1},g_{2},g_{3},g_{4} corresponding to \eta_{1},\eta_{2},\eta_{3},\eta_{4} . Then d_{3}=\begin{pmatrix}y_{2}&\alpha y_{1}&0&0\\ 0&0&x_{1}&x_{2}\\ x_{2}&0&-y_{2}&0\\ x_{1}&0&0&-\alpha y_{1}\\ 0&x_{2}&-y_{1}&0\\ 0&x_{1}&0&-y_{2}\end{pmatrix}. (4.7) A direct calculation shows that d_{2}d_{3}=0 . There is also a unique quartic syzygy among \eta_{1},\eta_{2},\eta_{3},\eta_{4} , namely \omega:=\alpha\eta_{1}y_{1}+\eta_{2}y_{2}+\eta_{3}x_{2}+\eta_{4}x_{1}. (4.8) Indeed, d_{3}(\omega)=\begin{pmatrix}\alpha(y_{2}y_{1}+y_{1}y_{2})\\ x_{1}x_{2}+x_{2}x_{1}\\ \alpha x_{2}y_{1}-y_{2}x_{2}\\ \alpha x_{1}y_{1}-\alpha y_{1}x_{1}\\ x_{2}y_{2}-y_{1}x_{2}\\ x_{1}y_{2}-y_{2}x_{1}\end{pmatrix}=0 in F_{2} by the defining relations of B . Let F_{4}=B(-4) with basis h and define d_{4}=\begin{pmatrix}\alpha y_{1}\\ y_{2}\\ x_{2}\\ x_{1}\end{pmatrix}. (4.9) Then d_{3}d_{4}=0 . Theorem 4.1 . Let B=B_{K}(\alpha) with \alpha\in\Bbbk^{\times} . Then the sequence 0\longrightarrow B(-4)\xrightarrow{d_{4}}B(-3)^{4}\xrightarrow{d_{3}}B(-2)^{6}\xrightarrow{d_{2}}B(-1)^{4}\xrightarrow{d_{1}}B\longrightarrow k_{B}\longrightarrow 0 is the minimal graded free resolution of the trivial right module k_{B} . In particular, \beta_{0,0}^{B}(k_{B})=1,\hskip 9.24994pt\beta_{1,1}^{B}(k_{B})=4,\hskip 9.24994pt\beta_{2,2}^{B}(k_{B})=6,\hskip 9.24994pt\beta_{3,3}^{B}(k_{B})=4,\hskip 9.24994pt\beta_{4,4}^{B}(k_{B})=1. Proof. We have already checked that d_{1}d_{2}=0 , d_{2}d_{3}=0 and d_{3}d_{4}=0 , so the displayed sequence is a complex. Minimality is immediate, since every matrix entry lies in B_{\geq 1} . Let V=\Bbbk x_{1}\oplus\Bbbk x_{2}\oplus\Bbbk y_{1}\oplus\Bbbk y_{2}, and let R\subseteq V^{\otimes 2} be the six-dimensional space spanned by the quadratic relations r_{1},\dots,r_{6} . Then B=T(V)/(R) and the first two maps of the displayed complex are exactly the first two maps of the right Koszul complex of B with respect to the chosen bases of V and R ; see [ 8 , Chs. 2 and 3] . Since B is Artin-Schelter regular of type (14641) , it is Koszul by Proposition 3.1 and therefore the right Koszul complex of B is exact [ 9 , 8 ] . Moreover, the graded Betti numbers of k_{B} are forced to be 1,4,6,4,1, so the third and fourth Koszul terms have ranks 4 and 1 , respectively. The four cubic syzygies \eta_{1},\eta_{2},\eta_{3},\eta_{4} are linearly independent and therefore provide a basis of the third Koszul space; similarly, the quartic syzygy \omega spans the fourth Koszul space. Consequently, the matrices ( 4.1 ), ( 4.2 ), ( 4.7 ) and ( 4.9 ) are precisely the coordinate form of the right Koszul complex of B . Hence the displayed complex is the minimal graded free resolution of k_{B} . The graded Betti numbers follow from the degree shifts of the free modules. ∎ Remark 4.2 . The explicit formulas for d_{3} and d_{4} show that the parameter \alpha survives all the way to the highest syzygies. In particular, the higher differentials retain concrete information about the defining double Ore data, rather than collapsing to a purely formal regularity pattern. 4.2. Natural cyclic right modules We now compute the minimal graded free resolutions of the natural cyclic right modules M_{x}=B/(x_{1},x_{2})B\text{ and }M_{y}=B/(y_{1},y_{2})B. Proposition 4.3 . The sequence 0\longrightarrow B(-2)\xrightarrow{\,d_{2}^{(x)}\,}B(-1)^{2}\xrightarrow{\,d_{1}^{(x)}\,}B\longrightarrow M_{x}\longrightarrow 0, where d_{1}^{(x)}(u,v)=ux_{1}+vx_{2},\ d_{2}^{(x)}(w)=(wx_{2},wx_{1}), is a minimal graded free resolution of M_{x} . In particular, \beta_{0,0}^{B}(M_{x})=1,\ \beta_{1,1}^{B}(M_{x})=2,\ \beta_{2,2}^{B}(M_{x})=1, and all other graded Betti numbers vanish. Proof. Clearly \operatorname{coker}d_{1}^{(x)}\cong M_{x} and d_{1}^{(x)}d_{2}^{(x)}(w)=w(x_{2}x_{1}+x_{1}x_{2})=0 by ( 3.1 ), so \operatorname{Im}d_{2}^{(x)}\subseteq\operatorname{Ker}d_{1}^{(x)} . The map d_{2}^{(x)} is injective because B is a domain. From the exact sequence 0\longrightarrow\operatorname{Ker}d_{1}^{(x)}\longrightarrow B(-1)^{2}\xrightarrow{\,d_{1}^{(x)}\,}B\longrightarrow M_{x}\longrightarrow 0 and the formulas H_{B}(t)=1/(1-t)^{4} , H_{M_{x}}(t)=1/(1-t)^{2} , we obtain H_{\operatorname{Ker}d_{1}^{(x)}}(t)=2tH_{B}(t)-H_{B}(t)+H_{M_{x}}(t)=\frac{t^{2}}{(1-t)^{4}}=H_{B(-2)}(t). Therefore \operatorname{Ker}d_{1}^{(x)}=\operatorname{Im}d_{2}^{(x)} and the sequence is exact. Minimality is immediate. ∎ Proposition 4.4 . The sequence 0\longrightarrow B(-2)\xrightarrow{\,d_{2}^{(y)}\,}B(-1)^{2}\xrightarrow{\,d_{1}^{(y)}\,}B\longrightarrow M_{y}\longrightarrow 0, where d_{1}^{(y)}(u,v)=uy_{1}+vy_{2},\hskip 18.49988ptd_{2}^{(y)}(w)=(wy_{2},wy_{1}), is a minimal graded free resolution of M_{y} . In particular, \beta_{0,0}^{B}(M_{y})=1,\hskip 18.49988pt\beta_{1,1}^{B}(M_{y})=2,\hskip 18.49988pt\beta_{2,2}^{B}(M_{y})=1, and all other graded Betti numbers vanish. Proof. The proof is identical to that of Proposition 4.3 , using the relation y_{2}y_{1}+y_{1}y_{2}=0 in place of x_{2}x_{1}+x_{1}x_{2}=0 . ∎ 5. Homological consequences The explicit computations above reveal a rigid but informative homological pattern for the pilot family K . The trivial right module k_{B} has the expected linear resolution of a four-dimensional regular algebra of type (14641) , while the natural cyclic modules M_{x} and M_{y} have linear resolutions of length 2 with graded Betti pattern (1,2,1) . Thus a single algebra simultaneously exhibits a four-dimensional regular scale encoded by k_{B} and a two-dimensional quadratic scale encoded by its most natural cyclic quotients. Theorem 5.1 . Let B=B_{K}(\alpha) with \alpha\in\Bbbk^{\times} . Then: (1) the trivial right module k_{B} has a minimal graded free resolution 0\longrightarrow B(-4)\longrightarrow B(-3)^{4}\longrightarrow B(-2)^{6}\longrightarrow B(-1)^{4}\longrightarrow B\longrightarrow k_{B}\longrightarrow 0; (2) the cyclic right modules M_{x} and M_{y} have minimal graded free resolutions 0\longrightarrow B(-2)\longrightarrow B(-1)^{2}\longrightarrow B\longrightarrow M_{x}\longrightarrow 0, and 0\longrightarrow B(-2)\longrightarrow B(-1)^{2}\longrightarrow B\longrightarrow M_{y}\longrightarrow 0; (3) the corresponding graded Betti patterns are (1,4,6,4,1),\hskip 18.49988pt(1,2,1),\hskip 18.49988pt(1,2,1). Proof. Part (1) is Theorem 4.1 , while part (2) is given by Propositions 4.3 and 4.4 . The graded Betti patterns follow immediately. ∎ Corollary 5.2 . The bigraded Poincaré series of these right B -modules are P_{k_{B}}^{B}(s,t)=1+4st+6s^{2}t^{2}+4s^{3}t^{3}+s^{4}t^{4}=(1+st)^{4}, and P_{M_{x}}^{B}(s,t)=P_{M_{y}}^{B}(s,t)=1+2st+s^{2}t^{2}=(1+st)^{2}. Proof. This is an immediate reformulation of the graded Betti numbers. ∎ Theorem 5.1 shows that the trivial module alone does not exhaust the homological information carried by the family. Although its Betti pattern is forced by regularity, the explicit matrices and the behavior of the natural cyclic quotients still reflect the defining parameter \alpha and the overlap structure of the PBW presentation. In particular, the pilot family provides a concrete instance in which the defining relations, the PBW basis and the minimal resolutions of natural modules can all be read from the same computation. Remark 5.3 . The explicit matrices obtained here are compatible with the general homological framework developed by Cao, Shen and Wang for graded double Ore extensions of Koszul algebras [ 4 ] . In the present pilot family, however, the PBW and overlap structure is simple enough to make all differentials visible directly in coordinates. The present calculations suggest that PBW and Gröbner-Shirshov methods provide an effective bridge between the defining relations of a trimmed double Ore extension and the explicit homological behavior of its graded modules. It would therefore be natural to carry out the same program for other genuinely double families in the Zhang-Zhang classification and to seek uniform matrix descriptions for the resulting minimal resolutions. 6. Future work The computations in this note suggest several natural directions for further investigation. A first one is to extend the present analysis to other genuinely double trimmed families in the Zhang–Zhang classification and compare the resulting minimal graded free resolutions. A second direction is to determine whether the natural cyclic right modules B/(x_{1},x_{2})B\ \text{and}\ B/(y_{1},y_{2})B continue to admit linear resolutions in wider classes of graded double Ore extensions. Finally, it would be interesting to seek uniform matrix descriptions for the higher differentials in terms of the defining DE-data, with the long-term goal of developing a broader explicit homological theory for trimmed graded double Ore extensions.