The algorithm engineering approach consists of a cycle of algorithm design, analysis, implementation, and experimental evaluation, with the aim of bridging the gap between theory and practice in the area of algorithms. This cycle of phases is driven by falsifiable hypotheses validated by experiments. Moreover, real-world instances often have direct impact on this cycle since they often expose modelling and analysis shortcomings. In the last fifteen years, this approach to algorithmic research has gained increasing attention. Prominent success stories in algorithm engineering include the algorithms library LEDA, the linear program solver CPLEX, the traveling salesman suite CONCORDE, speed-up techniques for shortest paths computation, as well as graph partitioning and the computation of Steiner trees. All these topics are driven by the need for efficient algorithms and libraries for problems that appear frequently in real-world applications. The seminar will cover all methodological aspects of algorithm engineering. Examples are:

The scientific method in algorithmics. One aspect of algorithm engineering is the scientific method, where research on algorithms is interpreted as in other disciplines such as physics and life sciences: the observation of a phenomenon that is not yet understood is investigated via falsifiable hypotheses as explanations of the phenomena, and experimental evaluations to test these hypotheses. That way not only emperical evidence on the behavior of algorithms is attained but also new theoretical insights are sought. Experimental algorithmics is already a core component of algorithm engineering from its very beginning. However, the design of reasonable experiments, the use of meaningful test instances, and reproducibility of experiments are still issues to be discussed in order to derive a common understanding and agree on a best practice.

The use of modern computer architecture. Exploiting the full potential of a modern computer poses many interesting new challenges for algorithm engineering: ever increasing parallelism, deep memory hierarchies, and heterogeneous architectures. Algorithms should be tailored to utilize multiple cores, but also access memory efficiently, taking into account issues such as data locality. Nowadays the use of GPUs, which are increasingly common in modern servers, is an important issue for efficient algorithm implementation. This is in particular interesting for frequently used and "classical" algorithms.

Certifying algorithms. An effective way to ensure correct results of algorithm implementations are certifying algorithms. The idea is to check each returned result for correctness using a simple checker. It then suffices to test or perhaps verify the checker. Making checking fast implies interesting algorithmic questions when checking is aided by certificates of correctness computed by the main algorithm.

Experiences from previous Dagstuhl Seminars show that the interaction between different scientific communities stimulates methodological discussions. This exchange is in particular important for neighboring scientific communities who typically meet at separate conferences. For this seminar, we plan to focus on web search, large graphs and social networks in order to also address the scientific WWW and Social Media community. In these fields, methods from algorithm engineering are applied. However, these scientists typically don’t publish at algorithm engineering conferences.

The aim of this seminar is to bring together researchers with different backgrounds in order to strengthen and foster collaborations and to identify key research directions for the future. In particular, the seminar is intended to foster the exchange between algorithm engineering and scientists from the web search community. While the dominant goal of the seminar is the exchange of current research developments and discussion of topical subjects, we also intend to bring algorithm engineering forward as a still evolving and expanding field in computer science. Therefore we will plan dedicated discussion sessions on methodological questions, but will also address research related issues like future DIMACS Implementation Challenges.

• Hannah Bast (Universität Freiburg, DE) [dblp]
• Jon Louis Bentley (Avaya - Basking Ridge, US) [dblp]
• Robert E. Bixby (Rice University - Houston, US) [dblp]
• Gerth Stølting Brodal (Aarhus University, DK) [dblp]
• Kevin Buchin (TU Eindhoven, NL) [dblp]
• Markus Chimani (Universität Osnabrück, DE) [dblp]
• Frank Dehne (Carleton University - Ottawa, CA) [dblp]
• Daniel Delling (Microsoft Corp. - Mountain View, US) [dblp]
• Julian Dibbelt (KIT - Karlsruher Institut für Technologie, DE) [dblp]
• Rudolf Fleischer (German University of Technology - Oman, OM) [dblp]
• Andrew V. Goldberg (Microsoft Corp. - Mountain View, US) [dblp]
• Holger H. Hoos (University of British Columbia - Vancouver, CA) [dblp]
• Falk Hüffner (TU Berlin, DE) [dblp]
• Giuseppe F. Italiano (University of Rome "Tor Vergata", IT) [dblp]
• David S. Johnson (New York, US) [dblp]
• Andrea Kappes (KIT - Karlsruher Institut für Technologie, DE) [dblp]
• Jyrki Katajainen (University of Copenhagen, DK) [dblp]
• Jürgen Lerner (Universität Konstanz, DE) [dblp]
• Kurt Mehlhorn (MPI für Informatik - Saarbrücken, DE) [dblp]
• Ulrich Carsten Meyer (Goethe-Universität - Frankfurt a. M., DE) [dblp]
• Friedhelm Meyer auf der Heide (Universität Paderborn, DE) [dblp]
• Henning Meyerhenke (KIT - Karlsruher Institut für Technologie, DE) [dblp]
• Shin-ichi Minato (Hokkaido University, JP) [dblp]
• Rolf H. Möhring (TU Berlin, DE) [dblp]
• Matthias Müller-Hannemann (Martin-Luther-Universität Halle-Wittenberg, DE) [dblp]
• Petra Mutzel (TU Dortmund, DE) [dblp]
• Patrick K. Nicholson (MPI für Informatik - Saarbrücken, DE) [dblp]
• Yoshio Okamoto (The University of Electro-Communications - Tokyo, JP) [dblp]
• Marina Papatriantafilou (Chalmers UT - Göteborg, SE) [dblp]
• Alejandro Salinger (Universität des Saarlandes, DE) [dblp]
• Peter Sanders (KIT - Karlsruher Institut für Technologie, DE) [dblp]
• Federico Santaroni (University of Rome "Tor Vergata", IT) [dblp]
• Sabine Storandt (Universität Freiburg, DE) [dblp]
• Dorothea Wagner (KIT - Karlsruher Institut für Technologie, DE) [dblp]
• Roger Wattenhofer (ETH Zürich, CH) [dblp]
• Renato Werneck (Microsoft Corp. - Mountain View, US) [dblp]
• Christos Zaroliagis (CTI & University of Patras, GR) [dblp]
• Liang Zhao (KIT - Karlsruher Institut für Technologie, DE) [dblp]
• Katharina A. Zweig (TU Kaiserslautern, DE) [dblp]