Analysis and Optimization of CHR Programs

Introduction. Constraint Handling Rules (CHR) [2] is a high-level, powerful, yet relatively simple “no box” CLP language, embedded in a host language, commonly Prolog. It is based on multi-headed committed-choice rules. Recent implementations of CHR consist of a compiler which translates a CHR program to host language code, and a run-time system implementing the constraint store. Originally, CHR was designed for rapid prototyping of user-defined constraint solvers. In the early years of CHR limited attention went to optimized compilation. As a consequence, the reference implementation of CHR [4] comprises a general compilation schema, with only a small number of optimizations. Currently, CHR is increasingly used as a general-purpose programming language in a wide range of applications. Therefore, performance becomes more important, and recently, more advanced compilation optimizations have been proposed [3]. Several implementations of CHR exist [10]. The K.U.Leuven CHR system [11] includes a state-of-the-art CHR compiler for popular Prolog systems like SWI-Prolog and XSB. The formulation of the refined operational semantics of CHR [1], and subsequently the formulation of the call-based refined operational semantics [5], have captured the essentials of current implementations on a formal level. Optimizations can now be defined formally and their correctness w.r.t. the operational semantics can be proved. The main goals of my research are to improve the practical usability of CHR, to allow a more declarative use of CHR, and to compile CHR programs to more efficient host language code. These goals are achieved by developping, implementing, and evaluating analyses and optimizations for CHR programs.