Introductory course on relation algebras, finite-dimensional cylindric algebras, and their interc
Introductory course on relation algebras, finite-dimensional cylindric algebras, and their interc
复制标题
关系代数、有限维圆柱代数及其相互关系入门课程
DOI:
--
复制
发表时间:
1990
期刊:
影响因子:
--
通讯作者:
R. Maddux
中科院分区:
文献类型:
--
作者:
R. Maddux
These are notes for a short course on relation algebras, nite-dimensional cylindric algebras, and their interconnections, delivered at the Conference on Alge-Relation algebras (RA's) are closely linked to cylindric algebras (CA's) and polyadic equality algebras (PEA's), and yet in certain ways they are quite diierent. The links and diierences allow for constructive interaction between these subjects. Relation algebras arise naturally by considering families of binary relations and certain operations on binary relations. The most crucial of these operations, relative multiplication (originally called composition by De Morgan D1856], D1864]), requires a reference to a third object in its deenition. Relative multiplication is an associative operation. The proof of this fact requires the examination of four objects. These informal observations are reeected in the following mathematical facts: relation algebras arise from those cylindric and polyadic algebras which are generated by their two-dimensional elements, which have at least three dimensions (so that relative multiplication is deenable), and which satisfy all the identities provable with four dimensions (so that relative multiplication turns out to be associative). The deenability of relative multiplication is a crucial feature which distinguishes cylindric and polyadic algebras of dimension 3 or more from those of lower dimension. Indeed, CA and PEA have undecidable equational theories just in case 3. Furthermore, if 3 then in each of these varieties the representable algebras form a proper subvariety which is not nitely axiomatizable, but this is not so whenever 2. Similarly, RA has an undecidable equational theory and the representable relation algebras form a nonnnitely based proper subvariety of RA. Therefore, throughout the rest of these notes we assume 3.