Complexity Of Infinite Domain Constraint Satisfaction

Introduces the universal-algebraic approach to classifying the computational complexity of constraint satisfaction problems.

Nowadays constraint satisfaction problems (CSPs) are ubiquitous in many different areas of computer science, from artificial intelligence and database systems to circuit design, network optimization, and theory of programming languages. Consequently, it is important to analyze and pinpoint the computational complexity of certain algorithmic tasks related to constraint satisfaction. The complexity-theoretic results of these tasks may have a direct impact on, for instance, the design and processing of database query languages, or strategies in data-mining, or the design and implementation of planners. This state-of-the-art survey contains the papers that were invited by the organizers after conclusion of an International Dagstuhl-Seminar on Complexity of Constraints, held in Dagstuhl Castle, Germany, in October 2006. A number of speakers were solicited to write surveys presenting the state of the art in their area of expertise. These contributions were peer-reviewed by experts in the field and revised before they were collated to the 9 papers of this volume. In addition, the volume contains a reprint of a survey by Kolaitis and Vardi on the logical approach to constraint satisfaction that first appeared in 'Finite Model Theory and its Applications', published by Springer in 2007.

In this thesis we study the computational complexity of MinCSP - an optimization version of the Constraint Satisfaction Problem (CSP). The input to a MinCSP is a set of variables and constraints applied to these variables, and the goal is to assign values (from a fixed domain) to the variables while minimizing the solution cost, i.e. the number of unsatisfied constraints. We are specifically interested in MinCSP with infinite domains of values. Infinite-domain MinCSPs model fundamental optimization problems in computer science and are of particular relevance to artificial intelligence, especially temporal and spatial reasoning. The usual way to study computational complexity of CSPs is to restrict the types of constraints that can be used in the inputs, and either construct fast algorithms or prove lower bounds on the complexity of the resulting problems. The vast majority of interesting MinCSPs are NP-hard, so standard complexity-theoretic assumptions imply that we cannot find exact solutions to all inputs of these problems in polynomial time with respect to the input size. Hence, we need to relax at least one of the three requirements above, opting for either approximate solutions, solving some inputs, or using super-polynomial time. Parameterized algorithms exploits the latter two relaxations by identifying some common structure of the interesting inputs described by some parameter, and then allowing super-polynomial running times with respect to that parameter. Such algorithms are feasible for inputs of any size whenever the parameter value is small. For MinCSP, a natural parameter is optimal solution cost. We also study parameterized approximation algorithms, where the requirement for exact solutions is also relaxed. We present complete complexity classifications for several important classes of infinite-domain constraints. These are simple temporal constraints and interval constraints, which have notable applications in temporal reasoning in AI, linear equations over finite and infinite fields as well as some commutative rings (e.g., the rationals and the integers), which are of fundamental theoretical importance, and equality constraints, which are closely related to connectivity problems in undirected graphs and form the basis of studying first-order definable constraints over infinite domains. In all cases, we prove results as follows: we fix a (possibly infinite) set of allowed constraint types C, and for every finite subset of C, determine whether MinCSP(), i.e., MinCSP restricted to the constraint types in , is fixed-parameter tractable, i.e. solvable in f(k) · poly(n) time, where k is the parameter, n is the input size, and f is any function that depends solely on k. To rule out such algorithms, we prove lower bounds under standard assumptions of parameterized complexity. In all cases except simple temporal constraints, we also provide complete classifications for fixed-parameter time constant-factor approximation.

In this thesis we study the worst-case complexity ofconstraint satisfaction problems and some of its variants. We use methods from universal algebra: in particular, algebras of total functions and partial functions that are respectively known as clones and strong partial clones. The constraint satisfactionproblem parameterized by a set of relations ? (CSP(?)) is the following problem: given a set of variables restricted by a set of constraints based on the relations ?, is there an assignment to thevariables that satisfies all constraints? We refer to the set ? as aconstraint language. The inverse CSPproblem over ? (Inv-CSP(?)) asks the opposite: given a relation R, does there exist a CSP(?) instance with R as its set of models? When ? is a Boolean language, then we use the term SAT(?) instead of CSP(?) and Inv-SAT(?) instead of Inv-CSP(?). Fine-grained complexity is an approach in which we zoom inside a complexity class and classify theproblems in it based on their worst-case time complexities. We start by investigating the fine-grained complexity of NP-complete CSP(?) problems. An NP-complete CSP(?) problem is said to be easier than an NP-complete CSP(?) problem if the worst-case time complexity of CSP(?) is not higher thanthe worst-case time complexity of CSP(?). We first analyze the NP-complete SAT problems that are easier than monotone 1-in-3-SAT (which can be represented by SAT(R) for a certain relation R), and find out that there exists a continuum of such problems. For this, we use the connection between constraint languages and strong partial clones and exploit the fact that CSP(?) is easier than CSP(?) when the strong partial clone corresponding to ? contains the strong partial clone of ?. An NP-complete CSP(?) problem is said to be the easiest with respect to a variable domain D if it is easier than any other NP-complete CSP(?) problem of that domain. We show that for every finite domain there exists an easiest NP-complete problem for the ultraconservative CSP(?) problems. An ultraconservative CSP(?) is a special class of CSP problems where the constraint language containsall unary relations. We additionally show that no NP-complete CSP(?) problem can be solved insub-exponential time (i.e. in2^o(n) time where n is the number of variables) given that theexponentialtime hypothesisis true. Moving to classical complexity, we show that for any Boolean constraint language ?, Inv-SAT(?) is either in P or it is coNP-complete. This is a generalization of an earlier dichotomy result, which was only known to be true for ultraconservative constraint languages. We show that Inv-SAT(?) is coNP-complete if and only if the clone corresponding to ? contains essentially unary functions only. For arbitrary finite domains our results are not conclusive, but we manage to prove that theinversek-coloring problem is coNP-complete for each k>2. We exploit weak bases to prove many of theseresults. A weak base of a clone C is a constraint language that corresponds to the largest strong partia clone that contains C. It is known that for many decision problems X(?) that are parameterized bya constraint language ?(such as Inv-SAT), there are strong connections between the complexity of X(?) and weak bases. This fact can be exploited to achieve general complexity results. The Boolean domain is well-suited for this approach since we have a fairly good understanding of Boolean weak bases. In the final result of this thesis, we investigate the relationships between the weak bases in the Boolean domain based on their strong partial clones and completely classify them according to the setinclusion. To avoid a tedious case analysis, we introduce a technique that allows us to discard a largenumber of cases from further investigation.

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

This book constitutes the proceedings of the 17th International Conference on Relational and Algebraic Methods in Computer Science, RAMiCS 2018, held in Groningen, The Netherlands, in October/November 2018. The 21 full papers and 1 invited paper presented together with 2 invited abstracts and 1 abstract of a tutorial were carefully selected from 31 submissions. The papers are organized in the following topics: Theoretical foundations; reasoning about computations and programs; and applications and tools.

This book constitutes the refereed proceedings of the 9th International Joint Conference on Automated Reasoning, IJCAR 2018, held in Oxford, United Kingdom, in July 2018, as part of the Federated Logic Conference, FLoC 2018. In 2018, IJCAR unites CADE, TABLEAUX, and FroCoS, the International Symposium on Frontiers of Combining Systems, and, for the fourth time, is part of the Federated Logic Conference. The 38 revised full research papers and 8 system descriptions presented together with two invited talks were carefully reviewed and selected from 108 submissions. The papers focus on topics such as logics, deductive systems, proof-search methods, theorem proving, model checking, verification, formal methods, and program analysis.

The two-volume set LNCS 5125 and LNCS 5126 constitutes the refereed proceedings of the 35th International Colloquium on Automata, Languages and Programming, ICALP 2008, held in Reykjavik, Iceland, in July 2008. The 126 revised full papers presented together with 4 invited lectures were carefully reviewed and selected from a total of 407 submissions. The papers are grouped in three major tracks on algorithms, automata, complexity and games, on logic, semantics, and theory of programming, and on security and cryptography foundations. LNCS 5126 contains 56 contributions of track B and track C selected from 208 submissions and 2 invited lectures. The papers for track B are organized in topical sections on bounds, distributed computation, real-time and probabilistic systems, logic and complexity, words and trees, nonstandard models of computation, reasoning about computation, and verification. The papers of track C cover topics in security and cryptography such as theory, secure computation, two-party protocols and zero-knowledge, encryption with special properties/quantum cryptography, various types of hashing, as well as public-key cryptography and authentication.