Description

Leader of the group Networks and Optimization: Guido Schäfer.

In today’s society, complex systems surround us. From transport and traffic, to behavioral economics and operations management, real-world applications often demand that we identify simple, optimal solutions among a huge set of possibilities. Our research group Networks and Optimization (N&O) does fundamental research to tackle such challenging optimization problems. 

We develop algorithmic methods to solve complex optimization problems efficiently. Our research provides efficient algorithms to some of the most challenging problems, for example, in planning, scheduling and routing. To come up with the best optimization algorithms, we combine and extend techniques from different disciplines in mathematics and computer science. 

N&O covers a broad spectrum of optimization aspects. Our expertise ranges from discrete to continuous optimization and applies to centralized and decentralized settings. We focus on both problem-specific methods and universal toolkits to solve different types of optimization problems. The key in our investigations is to understand and exploit combinatorial structures, such as graphs, networks, lattices and matroids. Our research is of high scientific impact and contributes to various fields.

In several cooperations with industry partners, the algorithmic techniques that we develop in our group have proven useful to solve complex real-world problems. We are always interested in new algorithmic challenges arising in real-world applications and are open to new cooperations.

 

Watch our group video to get a glimpse of our activities.

 

 

 

READ MORE...

Vacancies

News

Current events

N&O seminar: Jan-Hendrik Lange (Max-Planck-Institut für Informatik)

  • 2019-01-23T14:00:00+01:00
  • 2019-01-23T15:00:00+01:00
January 23 Wednesday

Start: 2019-01-23 14:00:00+01:00 End: 2019-01-23 15:00:00+01:00

CWI, Lecture room L016

 

Everyone is welcome to attend this N&O lecture:

Speaker: Jan-Hendrik Lange (Max-Planck-Institut für Informatik)
Title: Combinatorial persistency criteria for multicut

The multicut problem (aka weighted correlation clustering) is a combinatorial optimization problem used to model, for instance, 2D- and 3D-segmentation tasks in computer vision. Despite its NP-hardness in the worst case, practical instances often exhibit substantial structural simplicity. This is indicated by high quality heuristic solutions and LP relaxations that are fairly tight. In this talk, we present methods that exploit these observations for the purpose of optimization. We introduce combinatorial persistency criteria to efficiently compute partial optimality guarantees on practical instances. Our techniques range from elementary criteria that can be checked enumeratively in a fast preprocessing step to more expensive criteria that incorporate both fast primal and dual heuristics. We demonstrate the effectiveness of our methods on common benchmarks as well as large biomedical 3D-segmentation instances.

Open Access: turning promises into reality

  • 2019-01-25T14:30:00+01:00
  • 2019-01-25T23:59:59+01:00
January 25 Friday

Start: 2019-01-25 14:30:00+01:00 End: 2019-01-25 23:59:59+01:00

Turing room, Science Park Congress Centre, Amsterdam

CWI organizes a meeting concerning Open Access aimed for a broad audience. It will be devoted to a discussion of Plan S, https://www.scienceeurope.org/wp-content/uploads/2018/09/Plan_S.pdf , an initiative for Open Access publishing, launched in September 2018, and supported by an international consortium of research funders, including NWO, and to two radical initiatives, SciPost, https://scipost.org/, and Open Research Central, https://openresearchcentral.org/ .

Programme 14:30  Prof. Jos Baeten (CWI, Amsterdam and ILLC, University of Amsterdam) Title: Implementing Plan-S

15:00  Prof. Jean-Sébastien Caux (Institute of Physics, University of Amsterdam) Title: Beyond Open Access: SciPost

15:45  Vitek Tracz (F1000, London, UK) Title: Beyond Open Access: Open Research Central - a model for universal platform for reporting research results

16:30 Discussion

17:00 Drinks

Please register (free of charge) by sending an email to: susanne.van.dam@cwi.nl

 

Early bird registration NMC 2019: deadline 31 December 2019

  • 2019-04-23T09:00:00+02:00
  • 2019-04-24T18:00:00+02:00
April 23 Tuesday

Start: 2019-04-23 09:00:00+02:00 End: 2019-04-24 18:00:00+02:00

Congress Centre Koningshof, Veldhoven

On April 23-24, 2019, the Dutch Mathematical Congress (NMC 2019) will be held in the Congress Centre Koningshof  (located in Veldhoven, close to Eindhoven).
  
Given the success of the first "NMC new style" earlier this year, we hope that in 2019 there will be even more participants. Once again we have compiled a very attractive programme with the following ingredients:
Plenary lectures by Barbara Gentz, Jan Hesthaven, Gábor Lugosi and Anke van Zuylen
Parallel sessions by the Gravitation programmes Networks and Quantum Software Consortium, and of pairs of clusters inspired by the new topics in the “Sectorplan”
A strategy session on the cluster evaluation and the “Sectorplan”
Speed dates with industry
Poster prize, KWG prize for PhD students, the Stieltjes prize and the new “innovation” prize for MSc students
Social programme with dinner and dance performance based on the ‘spurs of Lehmer’

The "early bird" registration period, giving you a discount of € 50,-,  ends December 31st, so register now!  
 
You can find more information and a link to the registration form.
 
With regards,
On behalf of the organising committee,
  
Evgeny Verbitsky (chair)

Members

Associated Members

Publications

Current projects with external funding

  • Continuous Methods in Discrete Optimization
  • Verbeteren van de efficiency en prestatie van logistieke processen in de binnenvaart
  • Wiskundecluster DIAMANT
  • Approximation Algorithms, Quantum Information and Semidefinite Optimization (AQSO)
  • Mixed-Integer Non-Linear Optimisation Applications (MINOA)
  • Polynomial Optimization, Efficiency through Moments and Algebra (POEMA)
  • Vóórkomen en voorkómen van incidenten op het spoor (PPS Prorail)
  • Towards a Quantitative Theory of Integer Programming (QIP)

Related partners

  • Alma Mater Studiorum-Universita di Bologna
  • Alpen-Adria-Universität Klagenfurt
  • CNR Pisa
  • CNRS
  • CTVrede
  • IBM
  • INRIA
  • Prorail
  • Rheinische Friedrich-Wilhelmus Universitaet Bonn
  • Technische Universität Dortmund
  • Tilburg University
  • Tromsø, Norway
  • Universita degli Studi di Firenze
  • Universität Konstanz
  • University of Birmingham