You are here: Home News Current News Press Review

Press Review

Please visit our press archive to view older press reports.

Publication in "Operations Research"

Our study "The Constrained Reliable Shortest Path Problem in Stochastic Time-Dependent Networks" has been accepted for publication in the journal "Operations Research" (VHB-JOURQUAL: A+) .

Authors: Matthias Ruß, Gunther Gust, Dirk Neumann

The constrained reliable shortest path problem in stochastic time-dependent networks (CRSP-STD) extends the reliable, the time-dependent and the constrained shortest path problem. In the CRSP-STD, shortest paths need to ensure on-time arrival with a given probability and additionally satisfy constraints on timedependent weights. Examples of such time-dependent weights in transport networks include time-varying congestion charges or dynamic fees for shared vehicles. If weights decrease over time, waiting at nodes can be beneficial and is therefore allowed in our problem formulation. Travel times are modeled as time-dependent random variables and assumed to satisfy stochastic FIFO. We introduce a weak form of this condition to extend applicability to networks with scheduled connections, e. g., public transport. To solve the CRSP-STD, we define essential paths, a subset of optimal paths. Essential paths have two main properties: firstly, they cover all non-dominated combinations of worst-case weights that occur in the set of optimal paths, secondly, their subpaths are non-dominated, which can be used for pruning. Multiple properties of essential paths are exploited in our exact solution method, which extends multi-objective A* search. Runtime complexity is analyzed in theory and in numerical experiments, which show that key elements of our solution method effectively improve runtime performance.

New publication in "Journal of Operations Management"

Our study "Exploratory data science for discovery and ex-ante assessment of operational policies: Insights from vehicle sharing" has been accepted for publication in the Journal of Operations Management.

Authors: Tobias BrandtOliver Dlugosch

The proliferation of mobile devices and the emergence of the Internet of Things are leading to an unprecedented availability of operational data. In this paper, we study how leveraging this data in conjunction with data science methods can help researchers and practitioners in the development and evaluation of new operational policies. Specifically, we introduce a two-stage framework for exploratory data science consisting of a policy identification stage and an ex-ante policy assessment stage. We apply the framework to the context of free-floating carsharing – a novel mobility service that is an example for data-rich smart city services. Through data exploration, we identify a novel preventive user-based relocation policy and provide an ex-ante assessment regarding the feasibility of its implementation. We discuss practical implications of our approach and results for shared-mobility providers as well as the relationship between data science and operations management research.

Two Master's Theses supervised by our Chair were awarded with prizes

The master's theses by Katharina Baur ("Applied Machine Learning: Improving Real Estate Price Predictions with Natural Language Processing and Bayesian Optimization", supervised by Markus Rosenfelder) and Kathrin Leppert ("Auswirkungen des Terroranschlags am Breitscheidplatz: Eine Themen - und Sentimentanalyse deutscher Tweets", supervised by Bernhard Lutz) have been awarded with the Ralf-Bodo-Schmidt-Preis and the Friedrich-A.-Lutz-Preis respectively. Both prices amount to at least 1500€ and they are awarded once per year by the Faculty of Economics and Behavioral Sciences to outstanding student theses.



Student Thesis Accepted for Presentation at SKILL 2020

A Master's Thesis supervised by our chair was accepted for presentation at the German Students Conference for Informatics (SKILL 2020). The thesis "Applied Machine Learning: Improving Real Estate Price Predictions with Natural Language Processing and Bayesian Optimization" was written by Katharina Baur and supervised by Markus Rosenfelder. Congratulations!

New publication in Information Sciences

Our study "Negation Scope Detection for Sentiment Analysis: A Reinforcement Learning Framework for Replicating Human Interpretations" has been accepted for publication in the journal "Information Sciences".

Authors: Nicolas Pröllochs, Stefan Feuerriegel, Bernhard Lutz, and Dirk Neumann 



Textual materials represent a rich source of information for improving the decision-making of people, businesses and organizations. However, for natural language processing (NLP), it is difficult to correctly infer the meaning of narrative content in the presence of negations. The reason is that negations can be formulated both explicitly (e.g., by negation words such as ''not'') or implicitly (e.g., by expressions that invert meanings such as ''forbid'') and that their use is further domain-specific. Hence, NLP requires a dynamic learning framework for detecting negations and, to this end, we develop a reinforcement learning framework for this task. Formally, our approach takes document-level labels (e.g., sentiment scores) as input and then learns a negation policy based on the document-level labels. In this sense, our approach replicates human perceptions as provided by the document-level labels and achieves a superior prediction performance. Furthermore, it benefits from weak supervision; meaning that the need for granular and thus expensive word-level annotations, as in prior literature, is replaced by document-level annotations. In addition, we propose an approach to interpretability: by evaluating the state-action table, we yield a novel form of statistical inference that allows us to test which linguistic cues act as negations.