On \(d\)-orthogonality of the Sheffer systems associated to a convolution semigroup (Q557717)

The following question is investigated: Which \textit{I. M. Sheffer} polynomials [Duke Math. J. 3, 593--609 (1937; Zbl 0018.13601)] can be associated to a convolution semigroup of probability measures, usually induced by a stochastic process with stationary and independent increments? The main result with \(d\)-orthogonality is the following theorem: The associated \(d\)-orthogonal polynomials \((d\in\{2,3,\dots\})\) are characterized by the class of generating probability measures, which belongs to the natural exponential family with polynomial variance functions of exact degree \(2d-1\). This theorem extends some results of (classical) orthogonality [cp. e.g. \textit{W. Schoutens} and \textit{J. L. Teugels}, Commun. Stat., Stochastic Models 14, No.~1--2, 335--349 (1998; Zbl 0895.60050)]. Some new sets of martingales are pointed out, too. For each integer \(d \geq 2\) the polynomials with \((2d-1)\)-term recurrence relation for the families of positive stable processes are completely illustrated.