Lμ→Lν equiconvergence of spectral decompositions for Dirac system with Lϰ potential
We consider 1d-Dirac operator LP,U acting in H=(L2[0,π])2 \begin{gather*} \ell(\mathbf y) = B\mathbf y + P(x)\mathbf y,\qquad B = \begin{pmatrix}-i&0\\0&i\end{pmatrix},\\ P(x) = \begin{pmatrix}p_1(x)&p_2(x)\\ p_3(x)&p_4(x) \end{pmatrix},\qquad\mathbf y = \begin{pmatrix}y_1(x)\\ y_2(x)\end{pmatrix} \end{gather*} with arbitrary regular boundary conditions U(y)=0. The functions pj(x), 1≤j≤4, assumed to be complex valued and summable. Any regular 1d-Dirac operator of such kind has purely discrete spectrum {λn}n∈Z, λn=n+O(1) as ∣n∣→∞. We consider spectral decomposition SP,U(f)=limm→∞Sm,P,U(f) associated with operator LP,U: Sm,P,U(f)=∑∣n∣≤m[⟨f,z2n⟩y2n+⟨f,z2n+1⟩y2n+1]. Here yn --- eigen- and associated functions of LP,U and {zn}n∈Z --- biorthogonal in H system. Our main result claims that for every regular 1d-Dirac operator with p1=p4=0 and p2,p3∈Lϰ[0,π], ϰ∈(1,∞], and for every function f(x)=(f1(x)f2(x)), f1,f2∈Lμ[0,π], μ∈[1,∞], the equiconvergence ∥Sm,P,U(f)−Sm,0,U(f)∥Lν→0as m→∞ in the norm of the space (Lν[0,π])2, holds provided that 1/ϰ+1/μ−1/ν≤1 (with one exception ϰ=ν=∞, μ=1). In particular, in the case ϰ=μ=2, ν=∞ we obtain uniform pointwise equiconvergence on [0,π] of the series Sm,P,U(f) and Sm,0,U(f).