OR Days 2017

Over the years, the Swiss Operations Research Days have established themselves as a unique opportunity for researchers from the Swiss academic community, IBM Research and Operations Research professionals in Swiss enterprises and public organizations to learn about each other's work and interests.

The 15th Swiss Operations Research Days follow this tradition and bring researchers from Swiss universities, institutes of technology, and IBM Research together to exchange current research developments in Operations Research. The aim is not just to exchange knowledge, but to promote collaboration, initiate new projects and to have a god time with inspiring discussions.

The event will take place at the Universiy of Fribourg, from June 29 to June 30 2017, and is jointly organized by DS&OR and SVOR/ASRO and sponsored by SVOR/ASRO.

We invite members of IBM Research and Swiss OR researchers to submit abstracts about their recent research. We strongly encourage PhD students to present their result. The deadline for submission is June 14, 2017.

To register you need to be a member of SVOR or an IBM employee. The membership fee for SVOR is 80 CHF (free for students). For more information see the Registration page. For organizational reasons it is mandatory to register until at latest June 14, 2017.

This year we will have the pleasure to listen to the following plenary speakers:

Title: New Relaxtion-Based Approaches for Online and Large-Scale Optimization

Abstract: Optimization has been a key tool in various decision making processes for a long time. However, the type of problems we face has since changed considerably, with the advent of Big Data being a key driver of this change. A typical further challenge of increasing importance is the need to take decisions online in settings where data gets revealed over time. Within this loosely defined context of large-scale and online optimization, there are a multitude of problems which are hard to approach with existing techniques. In particular, designing solution techniques that are at the same time sufficiently fast to tackle even very large data sets, and deliver provably close-to-optimal solutions is a major challenge. In this talk, we show examples of how well-designed relaxations can achieve these goals in some relevant problem settings in online optimization and large-scale diversity maximization.

Title: Scaling up data mining techniques

Abstract: The overwhelming amount of available data poses a challenge for current data mining techniques. For large-scale data sets, the techniques either require extensive running time or tend to deliver unsatisfactory solutions. This talk provides an overview of widely-used classification techniques and compares them in terms of running time and performance. A novel scaling approach is proposed that can be applied to a wide range of classification and also clustering techniques. Computational results demonstrate that the scaling approach drastically reduces the running time of leading classification and clustering techniques with minor and often no loss in performance.

Title: The Swiss Data Science Center - A national center for Data Science

Abstract: The Swiss Data Science Center (www.datascience.ch) is a joint venture between EPFL and ETH Zurich, with offices in Lausanne and Zurich. Its mission is to accelerate the adoption of data science and machine learning techniques within academic disciplines of the ETH Domain, the Swiss academic community at large, and the industrial sector. In particular, it addresses the gap between those who create data, those who develop data analytics and systems, and those who could potentially extract value from it. The center is composed of a large multi-disciplinary team of data and computer scientists, and experts in select domains, with offices in Lausanne and Zurich. In this presentation, I will briefly describe the motivation and mission of the Swiss Data Science Center. I will then explain the vision and roadmap of the Swiss Open Insights Factory (SOIF), our hosted platform for Open Data Science. I will illustrate the current status of its development with a short demo.

  • Dr. Karl Isler (Swiss International Air Lines Ltd.)

Title: A robust heuristic for capacity control with planned upgrades

Abstract: In airline inventory optimization one has to consider the possibility to upgrade passengers to a higher compartment when there is excess demand for lower compartments. The booking control resulting from the higher dimensional generalization of the currently applied DP algorithm suffers from the curse of dimensionality and is technically not feasible. We present a heuristic method, derived from the deterministic fluid model, where it is only necessary to solve one dimensional dynamic programs and which can therefore easily be implemented in the existing reservation systems.