On transformations and zeros of polynomials (Q1178591): Difference between revisions

This is a survey of recent results by the authors about the following problem. Determine linear transformations \(T\) of the space \(\pi_ n\) of \(n\)-th degree polynomials into itself such that if all zeros of \(P\in \pi_ n\) lie in an interval \((a,b)\), then all zeros of \(TP\) lie in \((c,d)\); for example, \((a,b)\) and \((c,d)\) could be \((0,\infty)\). It is sufficient to describe the action of \(T\) on a basis of \(\pi_ n\). Two typical results: If \(T\) maps \(x^ n\) onto \(q^{n^ 2}x^ n\) for some \(q\in(-1,+1)\), then polynomials \(P\) with all real zeros are mapped into \(TP\) with the same property (Laguerre, 1982). If \(T\) maps \((x)_ n\) ( Pochhammer symbol) onto \(x^ n\), then polynomials with all zeros \(> 0\) are mapped into \(TP\) with the same property. Maps \(T\) with the desired properties can be constructed by the use of biorthogonal polynomials. Many examples are given.