[Job-offers-cs] Postdoc position at CWI

[Pekka Orponen]

-------- Forwarded Message --------
Subject: [DMANET] Postdoc position at CWI
Date: Tue, 17 Dec 2019 08:48:02 +0100
From: Daniel Dadush <D.N.Dadush at cwi.nl>
To: dmanet at zpr.uni-koeln.de

Hi, I would like to advertise the following position:

Title: Postdoc at CWI on the subject of "Towards a Quantitative Theory 
of Integer Programming"
Offered by: Daniel Dadush

Description:

The Networks & Optimization group at CWI is looking for a postdoctoral 
researcher with a strong interest in the theory and practice of Integer 
Programming (IP). The research will be supported by the ERC grant 
"Towards a Quantitative Theory of Integer Programming" and will take 
place at CWI in Amsterdam. The goals of this research are (1) to develop 
a quantitative theory that can explain the effectiveness of the 
prevalent techniques used for solving IPs (e.g. branch & bound, cutting 
planes, diving heuristics, etc.), (2) to develop new and effective 
techniques for solving IPs and (3) to build new connections between the 
study of IP, theoretical computer science and optimization. The research 
projects are designed to be interdisciplinary and are expected to 
require combining techniques from various areas of optimization (first 
order methods, interior point methods, simplex algorithms), theoretical 
computer science (discrepancy theory, fixed parameter tractability, 
smoothed analysis) and geometry (convex geometry, geometry of Euclidean 
lattices). Some sample research questions include:

- Can one obtain non-trivial upper bounds on the size of branch & bound 
trees produced by standard branching rules?
- Is there a cutting plane based polynomial time approximation scheme 
for the metric traveling salesman problem? More generally, can one 
develop a cutting plane framework with provable convergence guarantees?
- Can one automate basic FPT algorithms using the branch & bound framework?
- Can one demonstrate the efficiency of the simplex method for solving 
related sequences of linear programs (as one would encounter within 
branch & bound or cutting plane generation)?
- Can one develop practical rounding heuristics using discrepancy 
minimization algorithms?
- Is there a 2^O(n) -time algorithm for general integer programming?

Applications are currently being accepted. The process will remain open 
until the position is filled. For more information, please consult the 
official advertisement here:

Postdoc: 
https://www.cwi.nl/jobs/vacancies/postdoc-on-the-subject-of-towards-a-quantitative-theory-of-integer-programming-1

Homepage: https://homepages.cwi.nl/~dadush/

Best regards,

Daniel Dadush