Exact Algorithms For Constraint Satisfaction Problems

The Boolean satisfiability problem (SAT) and its generalization to variables of higher arities - constraint satisfaction problems (CSP) - can arguably be called the most "natural" of all NP-complete problems. The present work is concerned with their algorithmic treatment. It consists of two parts. The first part investigates CSPs for which satisfiability follows from the famous Lovasz Local Lemma. Since its discovery in 1975 by Paul Erdos and Laszlo Lovasz, it has been known that CSPs without dense spots of interdependent constraints always admit a satisfying assignment. However, an iterative procedure to discover such an assignment was not available. We refine earlier attempts at making the Local Lemma algorithmic and present a polynomial time algorithm which is able to make almost all known applications constructive. In the second part, we leave behind the class of polynomial time tractable problems and instead investigate the randomized exponential time algorithm devised and analyzed by Uwe Schoning in 1999, which solves arbitrary clause satisfaction problems. Besides some new interesting perspectives on the algorithm, the main contribution of this part consists of a refinement of earlier approaches at derandomizing Schoning's algorithm. We present a deterministic variant which losslessly reaches the performance of the randomized original.

This seminal text of Computer Science, the most cited book on the subject, is now available for the first time in paperback. Constraint satisfaction is a decision problem that involves finite choices. It is ubiquitous. The goal is to find values for a set of variables that will satisfy a given set of constraints. It is the core of many applications in artificial intelligence, and has found its application in many areas, such as planning and scheduling. Because of its generality, most AI researchers should be able to benefit from having good knowledge of techniques in this field. Originally published in 1993, this now classic book was the first attempt to define the scope of constraint satisfaction. It covers both the theoretical and the implementation aspects of the subject. It provides a framework for studying this field, relates different research, and resolves ambiguity in a number of concepts and algorithms in the literature. This seminal text is arguably the most rigorous book in the field. All major concepts were defined in First Order Predicate Calculus. Concepts defined this way are precise and unambiguous.

Search is an important component of problem solving in artificial intelligence (AI) and, more generally, in computer science, engineering and operations research. Combinatorial optimization, decision analysis, game playing, learning, planning, pattern recognition, robotics and theorem proving are some of the areas in which search algbrithms playa key role. Less than a decade ago the conventional wisdom in artificial intelligence was that the best search algorithms had already been invented and the likelihood of finding new results in this area was very small. Since then many new insights and results have been obtained. For example, new algorithms for state space, AND/OR graph, and game tree search were discovered. Articles on new theoretical developments and experimental results on backtracking, heuristic search and constraint propaga tion were published. The relationships among various search and combinatorial algorithms in AI, Operations Research, and other fields were clarified. This volume brings together some of this recent work in a manner designed to be accessible to students and professionals interested in these new insights and developments.

Graphical models (e.g., Bayesian and constraint networks, influence diagrams, and Markov decision processes) have become a central paradigm for knowledge representation and reasoning in both artificial intelligence and computer science in general. These models are used to perform many reasoning tasks, such as scheduling, planning and learning, diagnosis and prediction, design, hardware and software verification, and bioinformatics. These problems can be stated as the formal tasks of constraint satisfaction and satisfiability, combinatorial optimization, and probabilistic inference. It is well known that the tasks are computationally hard, but research during the past three decades has yielded a variety of principles and techniques that significantly advanced the state of the art. In this book we provide comprehensive coverage of the primary exact algorithms for reasoning with such models. The main feature exploited by the algorithms is the model's graph. We present inference-based, message-passing schemes (e.g., variable-elimination) and search-based, conditioning schemes (e.g., cycle-cutset conditioning and AND/OR search). Each class possesses distinguished characteristics and in particular has different time vs. space behavior. We emphasize the dependence of both schemes on few graph parameters such as the treewidth, cycle-cutset, and (the pseudo-tree) height. We believe the principles outlined here would serve well in moving forward to approximation and anytime-based schemes. The target audience of this book is researchers and students in the artificial intelligence and machine learning area, and beyond.

Presents a novel form of a compendium that classifies an infinite number of problems by using a rule-based approach.

Graphical models (e.g., Bayesian and constraint networks, influence diagrams, and Markov decision processes) have become a central paradigm for knowledge representation and reasoning in both artificial intelligence and computer science in general. These models are used to perform many reasoning tasks, such as scheduling, planning and learning, diagnosis and prediction, design, hardware and software verification, and bioinformatics. These problems can be stated as the formal tasks of constraint satisfaction and satisfiability, combinatorial optimization, and probabilistic inference. It is well known that the tasks are computationally hard, but research during the past three decades has yielded a variety of principles and techniques that significantly advanced the state of the art. This book provides comprehensive coverage of the primary exact algorithms for reasoning with such models. The main feature exploited by the algorithms is the model's graph. We present inference-based, message-passing schemes (e.g., variable-elimination) and search-based, conditioning schemes (e.g., cycle-cutset conditioning and AND/OR search). Each class possesses distinguished characteristics and in particular has different time vs. space behavior. We emphasize the dependence of both schemes on few graph parameters such as the treewidth, cycle-cutset, and (the pseudo-tree) height. The new edition includes the notion of influence diagrams, which focus on sequential decision making under uncertainty. We believe the principles outlined in the book would serve well in moving forward to approximation and anytime-based schemes. The target audience of this book is researchers and students in the artificial intelligence and machine learning area, and beyond.

A Constraint Satisfaction Problem (CSP) consists of a set of variables, a domain of values for each variable and a set of constraints. The objective is to assign a value for each variable such that all constraints are satisfied. CSPs continue to receive increased attention because of both their high complexity and their omnipresence in academic, industrial and even real-life problems. This is why they are the subject of intense research in both artificial intelligence and operations research. This book introduces the classic CSP and details several extensions/improvements of both formalisms and techniques in order to tackle a large variety of problems. Consistency, flexible, dynamic, distributed and learning aspects are discussed and illustrated using simple examples such as the n-queen problem. Contents 1. Foundations of CSP. 2. Consistency Reinforcement Techniques. 3. CSP Solving Algorithms. 4. Search Heuristics. 5. Learning Techniques. 6. Maximal Constraint Satisfaction Problems. 7. Constraint Satisfaction and Optimization Problems. 8. Distibuted Constraint Satisfaction Problems. About the Authors Khaled Ghedira is the general managing director of the Tunis Science City in Tunisia, Professor at the University of Tunis, as well as the founding president of the Tunisian Association of Artificial Intelligence and the founding director of the SOIE research laboratory. His research areas include MAS, CSP, transport and production logistics, metaheuristics and security in M/E-government. He has led several national and international research projects, supervised 30 PhD theses and more than 50 Master’s theses, co-authored about 300 journal, conference and book research papers, written two text books on metaheuristics and production logistics and co-authored three others.

Many fundamental combinatorial problems, arising in such diverse fields as artificial intelligence, logic, graph theory, and linear algebra, can be formulated as Boolean constraint satisfaction problems (CSP). This book is devoted to the study of the complexity of such problems. The authors' goal is to develop a framework for classifying the complexity of Boolean CSP in a uniform way. In doing so, they bring out common themes underlying many concepts and results in both algorithms and complexity theory. The results and techniques presented here show that Boolean CSP provide an excellent framework for discovering and formally validating "global" inferences about the nature of computation.