Michael Pinsker

In this paper, we investigate structural implications arising from the condition that a given directed graph does not interpret, in the sense of primitive positive interpretation with parameters or orbits, every finite structure. Our results generalize several theorems from the literature and yield further algebraic invariance properties that must be satisfied in every such graph. Algebraic properties of this kind are tightly connected to the tractability of constraint satisfaction problems, and we obtain new such properties even for infinite countably categorical graphs. We balance these positive results by showing the existence of a countably categorical hypergraph that fails to interpret some finite structure, while still lacking some of the most essential algebraic invariance properties known to hold for finite structures.

Many fundamental mathematical structures, such as the rationals or the random graph, are homogeneous, meaning that local isomorphisms extend to global automorphisms. Such structures arise as limits of classes of finite structures and encode these classes in a single object. This viewpoint has proved fruitful in model theory, universal algebra, and computer science, with applications to constraint satisfaction, automata theory, and verification. Homogeneous structures have rich automorphism groups, which makes them interesting for topological dynamics. For many applications, however, automorphism groups do not store enough information about the homogeneous structure, and one must instead consider polymorphism clones. Universal algebra has recently achieved major results for polymorphism clones on finite structures, culminating in the 2017 resolution of the Feder–Vardi dichotomy conjecture. An analogous conjecture for homogeneous structures remains open despite growing structural insights.

For a set [Formula: see text] of finite tournaments, the [Formula: see text]-free orientation problem is the problem of deciding if a given finite undirected graph can be oriented in such a way that the resulting oriented graph does not contain any member of [Formula: see text]. Using the theory of smooth approximations, we give a new shorter proof of the complexity dichotomy for such problems obtained recently by Bodirsky and Guzmán-Pro. In fact, our approach yields a complexity dichotomy for a considerably larger class of computational problems, where one is given an undirected graph along with additional local constraints on the allowed orientations. Moreover, the border between tractable and hard problems is also described by a decidable algebraic condition.

We prove that relational structures admitting specific polymorphisms (namely, canonical pseudo-WNU operations of all arities [Formula: see text]) have low relational width. This implies a collapse of the bounded width hierarchy for numerous classes of infinite-domain constraint satisfaction problems (CSPs) studied in the literature. Moreover, we obtain a characterization of bounded width for first-order reducts of unary structures and a characterization of Monotone Monadic SNP (MMSNP) sentences that are equivalent to a Datalog program, answering a question posed by Bienvenu et al. In particular, the bounded width hierarchy collapses in those cases as well. Our results extend the scope of theorems of Barto and Kozik characterizing bounded width for finite structures and show the applicability of infinite-domain CSPs to other fields.

We introduce the novel machinery of smooth approximations to provide a systematic algebraic approach to the complexity of CSPs over finitely bounded homogeneous structures. We apply smooth approximations to confirm the CSP dichotomy conjecture for first-order reducts of the random tournament and to give new short proofs of the conjecture for various homogeneous graphs including the random graph (STOC’11, ICALP’16, JACM 2015, SICOMP 2019), and for expansions of the order of the rationals (STOC’08, JACM 2009). Apart from obtaining these dichotomy results, we show how our new proof technique allows one to unify and significantly simplify the previous results from the literature. For all but the last structure, we moreover characterize for the first time those CSPs that are solvable by local consistency methods, again using the same machinery.

We investigate the `local consistency implies global consistency' principle of strict width among structures within the scope of the Bodirsky-Pinsker dichotomy conjecture for infinite-domain Constraint Satisfaction Problems (CSPs). Our main result implies that for certain CSP templates within the scope of that conjecture, having bounded strict width has a concrete consequence on the expressive power of the template called implicational simplicity. This in turn yields an explicit bound on the relational width of the CSP, i.e., the amount of local consistency needed to ensure the satisfiability of any instance. Our result applies to first-order expansions of any homogeneous $k$-uniform hypergraph, but more generally to any CSP template under the assumption of finite duality and general abstract conditions mainly on its automorphism group. In particular, it overcomes the restriction to binary signatures in the pioneering work of Wrona.

The set of increasing functions on the rational numbers, equipped with the composition operation, naturally forms a topological semigroup with respect to the topology of pointwise convergence in which a sequence of increasing functions converges if and only if it is eventually constant at every argument. We develop new techniques to prove there is no other Polish topology turning this semigroup into a topological one, and show that previous techniques are insufficient for this matter.

We prove that relational structures admitting specific polymorphisms (namely, canonical pseudo-WNU operations of all arities $n \geq 3$) have low relational width. This implies a collapse of the bounded width hierarchy for numerous classes of infinite-domain CSPs studied in the literature. Moreover, we obtain a characterization of bounded width for first-order reducts of unary structures and a characterization of MMSNP sentences that are equivalent to a Datalog program, answering a question posed by Bienvenu, ten Cate, Lutz, and Wolter. In particular, the bounded width hierarchy collapses in those cases as well. Our results extend the scope of theorems of Barto and Kozik characterizing bounded width for finite structures, and show the applicability of infinite-domain CSPs to other fields.

A Constraint Satisfaction Problem (CSP) is a computational problem where we are given variables and constraints about them; the question is whether the variables can be assigned values such that all constraints are satisfied. We give an overview of the current state of research on CSPs where values for the variables and constraints are taken from a finitely bounded homogeneous structure which is fixed beforehand. We explain the main mathematical ideas so far, the three dilemmas they brought upon us, and what could be done to overcome them in order to obtain a satisfactory understanding of the computational complexity of such CSPs.