Genetic Programming Hyper-heuristics (GPHH) have been successfully used to evolve scheduling rules for Dynamic Workflow Scheduling (DWS) as well as other challenging combinatorial optimization problems. The method of sampling training instances has a significant impact on the generalization ability of GPHH, yet they are rarely addressed in existing research. This article aims to fill this gap by proposing a GPHH algorithm with a sampling strategy to thoroughly investigate the impact of six instance sampling strategies on algorithmic generalization, including one rotation strategy, three mini-batch strategies, and two hybrid strategies. Experiments across four scenarios with varying settings reveal that: (1) mini-batch with random sampling can outperform rotation in generalizing to unseen workflow scheduling problems under the same computational cost; (2) employing a hybrid strategy that combines rotation and mini-batch further enhances the generalization ability of GPHH; and (3) mini-batch and hybrid strategies can effectively enable heuristics trained on small-scale training instances generalizing well to large-scale unseen ones. These findings highlight the potential of mini-batch strategies in GPHH, offering improved generalization performance while maintaining diversity and suggesting promising avenues for further exploration in GPHH domains.

Embedded functional dependencies (eFDs) advance data management applications by data completeness and integrity requirements. We show that the discovery problem of eFDs is NP-complete, W[2]-complete in the output, and has a minimum solution space that is larger than the maximum solution space for functional dependencies. Nevertheless, we use novel data structures and search strategies to develop row-efficient, column-efficient, and hybrid algorithms for eFD discovery. Our experiments demonstrate that the algorithms scale well in terms of their design targets, and that ranking the eFDs by the number of redundant data values they cause can provide useful guidance in identifying meaningful eFDs for applications. Finally, we demonstrate the benefits of introducing completeness requirements and ranking by the number of redundant data values for approximate and genuine functional dependencies.

Cardinality constraints and functional dependencies can enforce complex business rules within database systems. As the interaction of these constraints is intricate on SQL data, data engineers and domain experts face the challenge of deciding which of the constraints are meaningful for the underlying application domain. We present a tool that computes data samples that perfectly summarize which of the constraints are currently perceived meaningful. It is demonstrated how the tool facilitates the interaction and understanding of data engineers and domain experts to help them separate meaningful from meaningless cardinality constraints and functional dependencies.

The data deluge is defined by increasing amounts of large data with increasing degree of uncertainty. In a recent response, probabilistic databases are receiving a great deal of interest from research and industry. One popular approach to probabilistic databases is to extend traditional relational database technology to handle uncertainty. In this approach probabilistic databases are probability distributions over a collection of possible worlds of relational databases. On the one hand, research has seen various efforts to extend query evaluation from relational to probabilistic databases. On the other hand, updates have not received much attention at all. In this paper we show that well-known syntactic normal form conditions capture probabilistic databases with desirable update behavior. Such behavior includes the absence of data redundancy, insertion, deletion, and modification anomalies. We further show that standard normalization procedures can be applied to standard representations of probabilistic databases to obtain database schemata that satisfy the normal form condition, and can thus be updated efficiently.

We investigate the implication problem for classes of data dependencies over SQL table definitions. Under Zaniolo's “no information” interpretation of null markers we establish an axiomatization and algorithms to decide the implication problem for the combined class of functional and multivalued dependencies in the presence of NOT NULL constraints. The resulting theory subsumes three previously orthogonal frameworks. We further show that the implication problem of this class is equivalent to that in a propositional fragment of Schaerf and Cadoli's [1995] family of para-consistent S-3 logics. In particular, S is the set of variables that correspond to attributes declared NOT NULL. We also show how our equivalences for multivalued dependencies can be extended to Delobel's class of full first-order hierarchical decompositions, and the equivalences for functional dependencies can be extended to arbitrary Boolean dependencies. These dualities allow us to transfer several findings from the propositional fragments to the corresponding classes of data dependencies, and vice versa. We show that our results also apply to Codd's null interpretation “value unknown at present”, but not to Imielinski's [1989] or-relations utilizing Levene and Loizou's weak possible world semantics [Levene and Loizou 1998]. Our findings establish NOT NULL constraints as an effective mechanism to balance not only the certainty in database relations but also the expressiveness with the efficiency of entailment relations. They also control the degree by which the implication of data dependencies over total relations is soundly approximated in SQL table definitions.

We study functional and multivalued dependencies over SQL tables with NOT NULL constraints. Under a no-information interpretation of null values we develop tools for reasoning. We further show that in the absence of NOT NULL constraints the associated implication problem is equivalent to that in propositional fragments of Priest's paraconsistent Logic of Paradox. Subsequently, we extend the equivalence to Boolean dependencies and to the presence of NOT NULL constraints using Schaerf and Cadoli's S-3 logics where S corresponds to the set of attributes declared NOT NULL. The findings also apply to Codd's interpretation "value at present unknown" utilizing a weak possible world semantics. Our results establish NOT NULL constraints as an effective mechanism to balance the expressiveness and tractability of consequence relations, and to control the degree by which the existing classical theory of data dependencies can be soundly approximated in practice.

Constraints are important for a variety of XML recommendations and applications. Consequently, there are numerous opportunities for advancing the treatment of XML semantics. In particular, suitable notions of keys will enhance XML's capabilities of modeling, managing and processing native XML data. However, the different ways of accessing and comparing XML elements make it challenging to balance expressiveness and tractability.

We review key constraints in the context of XML as introduced by Buneman et al. We demonstrate that: (1) one of the proposed inference rules is not sound in general, and (2) the inference rules are incomplete for XML key implication, even for nonempty sets of simple key paths. This shows, in contrast to earlier statements, that the axiomatizability of XML keys is still open, and efficient algorithms for deciding their implication still need to be developed. Solutions to these problems have a wide range of applications including consistency validation, XML schema design, data exchange and integration, consistent query answering, XML query optimization and rewriting, and indexing. In this article, we investigate the axiomatizability and implication problem for XML keys with nonempty sets of simple key paths. In particular, we propose a set of inference rules that is indeed sound and complete for the implication of such XML keys. We demonstrate that this fragment is robust by showing the duality of XML key implication to the reachability problem of fixed nodes in a suitable digraph. This enables us to develop a quadratic-time algorithm for deciding implication, and shows that reasoning about this XML key fragment is practically efficient. Therefore, XML applications can be unlocked effectively since they benefit not only from those XML keys specified explicitly by the data designer but also from those that are specified implicitly.

The study of integrity constraints has been identified as one of the major challenges in XML database research. The main difficulty is finding a balance between the expressiveness and the existence of automated reasoning tools for different classes of constraints.

Multi-valued depdendencies (MVDs) are an important class of relational constraints. We axiomatise MVDs in data models that support nested list types. In order to capture different data models at a time, an abstract approach based on nested attributes is taken. The set of subattributes of some fixed nested attribute carries the structure of a co-Heyting algebra. This enables us to generalise significant features of MVDs from the relational data model to the presence of lists. It is shown that an MVD is satisfied by some instance exactly when this instance can be decomposed without loss of information. The full power of the algebraic framework allows to provide a sound and complete set of inference rules for the finite implication of MVDs in the context of lists. The presence of the list operator calls for a new inference rule which is not required in the relational data model. Further differences become apparant when the minimality of the inference rules is investigated. The extension of the relational theory of MVDs to the presence of lists allows to specify more real-world constraints and increases therefore the number of application domains.