The Design And Implementation Of A CLP(R) Compiler

My graduate thesis is CLP(R) Compiler System, supported by Funding of Chinese Nature Science and 863 Project. This paper is a summary of my graduate study in the past three years. By unifying constraint satisfaction technique and Logic Programming method, Constraint Logic Programming, which not only maintains the natural and succinct attribute for expressing constraints satisfied problem of Logic Programming, but also solves these problem efficiently, makes constraint solving method for the particular domain instead of the unification function in Logic Programming. In December of 1996, ACM made CLP one of the 19 key research areas of computer science in the future. There are several criteria to classify Constraint Logic Programming and its system. The first one is assorted to single domain and multiple domains by their handling constraint domain. The second one is assorted to linear constraint and nonlinear constraint by their handling constraint categories. The third one is assorted to Interpreter System and Compiler System by their implementing function. So far we have not found in literature any similar CLP system combining the three features: multi-domain, non-linear constraints and compilation. The compiling technique, system implementing, optimizing strategies, implementing examples of a CLP system are discussed in this paper. The system is a multi-domain CLP system, based on interval arithmetic and constraints propagation as the unified constraint solving method for the discrete domain and real domain. This compiler can solve nonlinear constraints based on an extended WAM instruction set. This paper focus on CLP(R) ?the real domain CLP, extending the CLP(FD) previously developed in our research group.