Interpolation by polynomial functions of distributive lattices: a generalization of a theorem of R. L. Goodstein (Q2377308): Difference between revisions

Let \(L\) be a distributive lattice with least and greatest elements \(0\) and \(1\). Let \(f\) be a function \(\{0,1\}^n\to L\). A well-known result of Goodstein and Grätzer says that \(f\) can be interpolated by a lattice polynomial if and only if \(f\) is isotone. Moreover, such a polynomial function is unique. The present paper generalizes this theorem to functions \(f:\;D\to L\), where \(D\subset L^n\) is a set of the form \(\{a_1,b_1\}\times\dots\times\{a_n,b_n\}\) with \(a_i<b_i\). In this case, the interpolating polynomial function need not be unique, so the authors describe all such polynomial functions. They also present an application of their result in decision making.