Ian Miguel Knihy

The book features a comprehensive exploration of various topics in constraint programming, optimization, and heuristic search. It includes invited talks and research papers that cover state abstraction in real-time heuristic search, the integration of optimization with constraint programming, and the generation of constraint models. Key discussions involve DFS-tree based heuristic search, partial pattern databases, and the use of constraint databases for program verification. The text delves into generating implied Boolean constraints, reformulating constraint satisfaction problems for better scalability, and the relaxation of qualitative constraint networks. It also addresses dynamic domain abstraction, approximate model-based diagnosis, and the combination of perimeter search with pattern database abstractions. Further, it examines the efficient conversion of ground logic satisfiability to propositional logic, tailoring solver-independent constraint models, and meta-CSP models for optimal planning. The book discusses formalizing the abstraction process in model-based diagnosis, active learning in dynamic Bayesian networks, and boosting minimal unsatisfiable set extraction. Additionally, it analyzes map-based abstraction and refinement, the relationship between abstraction and complexity measures, and the integration of constraint programming with metaheuristics. The work emphasizes the importance of analogy discovery,