Finite State Relational Programs

We introduce a new class of programs, called Finite Relational Linear Programs (FRLP), capable of modeling simple data processing applications. We analyze these programs with respect to tradeoffs between features they possess including the problems of equivalence and optimization. Algorithmic solutions are given to the two latter problems and others. Extensions are discussed in which a (part of) algorithmic solvability is lost. A unifying concept for this and previous work is presented.