Homology of nilpotent subalgebras of the Lie superalgebra \(K(1,1)\). III (Q1810302)

The Witt algebra \(W\), i.e. the Lie algebra defined by generators \(e_i\), \(i\in{\mathbb Z}\), and relations \([e_i,e_j]=(j-i)e_{i+j}\), has a supersymmetric relative called \(K(1,1)\), which is defined by generators \(e_i\), \(i\in{\mathbb Z}\), which are even or odd according to the parity of \(i\), and relations \[ [e_{2i+1},e_{2j+1}]=e_{2i+2j+2} , [e_{2i},e_{2j+1}]=(2j-i+1)e_{2i+2j+1} , [e_{2i},e_{2j}]=2(j-i)e_{2i+2j}. \] The homology of the subalgebra \(L_1(W)= \text{ span} \{e_i |i\geq 1\}\) has been calculated by Goncharova and plays a central rôle in the cohomology theory of the Lie algebra of vector fields on the circle or on the line. In the article under review, the author continues his study of the homology of the corresponding subalgebra \(L_n:= \text{ span} \{e_i |i\geq n\}\leq K(1,1)\). For Part II, see Russ. Math. Surv. 55, 1154-1156 (2000); translation from Usp. Mat. Nauk 55, No. 6, 143-144 (2000; Zbl 1026.17024)]. Denote by \(C_*(L_n)\) the Chevalley-Eilenberg complex of the Lie superalgebra \(L_n\), \(Z_*(L_n)\) the subspace of cocycles and \(H_*(L_n)\) the homology with trivial coefficients. The \({\mathbb Z}\)-grading on \(K(1,1)\) and \(L_n\) induces a \({\mathbb Z}\)-grading on \(C_*(L_n)\), and we denote by \(C_*^{(m)}(L_n)\), \(Z_*^{(m)}(L_n)\) and \(H_*^{(m)}(L_n)\) the corresponding spaces of internal degree \(m\). By explicitly calculating the dimensions of images and kernels of the boundary operator, the author calculates the dimension of \(H_2^{(m)}(L_n)\). The computations are quite long and combinatorially involved. As an application, \(H^2(L_1,L_1)\) is computed in order to describe infinitesimal deformations of \(L_1\). The computation is performed via the spectral sequence induced by the filtration of the cochain complex associated to the internal degree. \(H^2(L_1,L_1)\) turns out to be of dimension 3, but these generators of non-trivial infinitesimal deformations are shown to be obstructed, so that they do not lead to (full) formal deformations. In this sense, \(L_1\) is rigid.