No abstract

Dynamic Programming (DP) over tree decompositions is a well-established method to solve problems -that are in general NP-hard -efficiently for instances of small treewidth. Experience shows that (i) heuristically computing a tree decomposition has negligible runtime compared to the DP step; and (ii) DP algorithms exhibit a high variance in runtime when using different tree decompositions; in fact, given an instance of the problem at hand, even decompositions of the same width might yield extremely diverging runtimes. We thus propose here a novel and general method that is based on selection of the best decomposition from an available pool of heuristically generated ones. For this purpose, we require machine learning techniques that provide automated selection based on features of the decomposition rather than on the actual problem instance. Thus, one main contribution of this work is to propose novel features for tree decompositions. Moreover, we report on extensive experiments in different problem domains which show a significant speedup when choosing the tree decomposition according to this concept over simply using an arbitrary one of the same width.

The notion of secure sets is a rather new concept in the area of graph theory. Applied to social network analysis, the goal is to identify groups of entities that can repel any attack or influence from the outside. In this article, we tackle this problem by utilizing Answer Set Programming (ASP). It is known that verifying whether a set is secure in a graph is already co-NP-hard. Therefore, the problem of enumerating all secure sets is challenging for ASP and its systems. In particular, encodings for this problem seem to require disjunction and also recursive aggregates. Here, we provide such encodings and analyse their performance using the Clingo system. Furthermore, we study several problem variants, including multiple secure or insecure sets, and weighted graphs.

Answer Set Programming (ASP) is a powerful declarative programming paradigm that has been successfully applied to many different domains. Recently, ASP has also proved successful for hard optimization problems like course timetabling and travel allotment. In this paper, we approach another important task, namely, the shift design problem, aiming at an alignment of a minimum number of shifts in order to meet required numbers of employees (which typically vary for different time periods) in such a way that over-and understaffing is minimized. We provide an ASP encoding of the shift design problem, which, to the best of our knowledge, has not been addressed by ASP yet. Our experimental results demonstrate that ASP is capable of improving the best known solutions to some benchmark problems. Other instances remain challenging and make the shift design problem an interesting benchmark for ASP-based optimization methods.

No abstract