Computing and using minimal polynomials (Q2307625)

This paper describes theoretical and practical aspects of the computation of minimal polynomials. Minimal polynomials have some advantages over characteristic polynomials, such as generalizing to families of pairwise commuting endomorphisms of finite-dimensional vector spaces. The authors are leaders in the field, and their previous work provides a foundation for these investigations. A significant aspect of this paper is the development of a modular approach to computing minimal polynomials of elements of an affine \(\mathbb{Q}\)-algebra that is guaranteed to produce the correct minimal polynomial (rather than one that has high probability of being correct). The algorithms given in this paper are implemented in CoCoA and CoCoALib, and the authors provide some timing comparisons with Singular and Macaulay2 that show that the implementations described here are highly efficient; however, these timings are somewhat wanting as the authors note that they are not experts in either Singular or Macaulay2, and thus may not have used the most optimal implementations available. Moreover, the authors were unable to obtain results from Macaulay2 in a reasonable timeframe. The paper is an enjoyable read, well-organized and well-written, with numerous examples illustrating the practical aspects of the research.