On a structure formula for classical \(q\)-orthogonal polynomials (Q5948568)

A short proof is given for the fact that any solution of the \(q\)-difference equation NEWLINE\[NEWLINE\sigma(x)D_qD_{1/q}y(x)+\tau(x)D_qy(x)+\lambda_{q,n}y(x)=0,NEWLINE\]NEWLINE where \(\sigma(x)=ax^2+bx+c\), \(\tau(x)=dx+e\), for some real numbers \(a,b,c,d,e\), \(D_q\) is the usual \(q\)-difference operator defined by \(D_qf(x):=(f(qx)-f(x))/(q-1)x\), and \(\lambda_{q,n}\) is (enforcedly) equal to \(-a[n]_{1/q} [n-1]_q-d[n]_q\) with \([n]_q:=(1-q^n) /(1-q)\), satisfies the ``structure formula'' NEWLINE\[NEWLINE\sigma(x)D_{1/q}P_n(x)=\alpha_nP_{n+1}(x)+\beta_nP_n(x)+\gamma_nP_{n-1} (x),NEWLINE\]NEWLINE with explicitly known real numbers \(\alpha_n,\beta_n,\gamma_n\). The proof is based on the fact that any solution of the \(q\)-differene equation satisfies a three-term recurrence NEWLINE\[NEWLINEP_{n+1}(x)=(A_nx+B_n)P_n(x)-C_{n}P_{n-1}(x)NEWLINE\]NEWLINE (and is therefore a family of orthogonal polynomials). The authors mention that an independent proof of the same fact was given by Steffen Häcker in his Ph.D. thesis ``Polynomiale Eigenwertprobleme zweiter Ordnung mit Hahnschen \(q\)-Operatoren,'' using a different approach.