Programming constraint services

Más información sobre el libro

Constraint Programming is an effective approach for modeling and solving combinatorial problems, leveraging techniques from Artificial Intelligence, Logic Programming, and Operations Research. Central to this method are constraint propagation and heuristic search. It abstracts problem-solving into a reusable constraint engine and a declarative program that models the problem, with the engine implementing necessary propagation and search algorithms. This engine can be integrated as a library in general-purpose programming languages (like C++), as an extension of existing languages (like Prolog), or as a standalone system with its own language. The focus here is on the architecture and implementation of constraint engines, introducing a new concurrent architecture that surpasses the sequential architecture of Prolog. This concurrent method employs copying and recomputation instead of sequential search with trailing and backtracking, allowing for any search strategy and simplifying the implementation of constraint propagation algorithms. Additionally, the book explores a generalization of this concurrent architecture that supports propagation-preserving combinators, such as deep guard combinators for negation, disjunction, implication, and recitation of constraint propagators, which are not feasible within Prolog’s framework. This innovative approach utilizes a reflective encapsulation primitive to achieve these capabilities.