Coinductive Constraint Logic Programming

Constraint logic programming (CLP) has been proposed as a declarative paradigm for merging constraint solving and logic programming. Recently, coinductive logic programming has been proposed as a powerful extension of logic programming for handling (rational) infinite objects and reasoning about their properties. Coinductive logic programming does not include constraints while CLP’s declarative semantics is given in terms of a least fixed-point (i.e., it is inductive) and cannot directly support reasoning about (rational) infinite objects and their properties. In this paper we combine constraint logic programming and coinduction to obtain co-constraint logic programming (co-CLP for brevity). We present the declarative semantics of co-CLP in terms of a greatest fixed-point and its operational semantics based on the coinductive hypothesis rule. We prove the equivalence of these two semantics for programs with rational terms.