Optimization Algorithm Developer Position @ Gurobi is available

 Gurobi Optimization is a leading provider of software libraries for 
solving mathematical programming problems. Founded in 2008
 by three of the world's leading experts in computational optimization, 
Gurobi now provides software libraries to hundreds of commercial
 clients in a variety of industries. As a result of our rapid growth,
 we're looking to expand our development team.

Job functions

• Develop and enhance algorithms for solving mathematical 
• programming problems (mixed integer programming and/or
•  linear, quadratic, and second-order cone programming).
• Work closely with a small team of highly skilled software 
• developers. The position will require significant collaborative
•  work, as well as significant independent work.
• Provide internal support, as needed, for our product support 
• and marketing teams.
• Present new product features and capabilities at technical 
• conferences, and interact with product users at such events.

Qualifications

• PhD in Operations Research, Computer Science, or similar
•  discipline.
• Three years experience in developing high performance
•  software with a significant mathematical component. To be
•  successful in this position, you must be skilled in 
• understanding, implementing, and extending complex 
• mathematical algorithms.

Location

• We are flexible on work location for this position,
•  since Gurobi is a highly distributed company with 
• employees in various locations across the USA.

Additional requirement for submitting an application

• To allow us to limit our attention to serious candidates only,
   we request that you implement and submit a small 
  subroutine to perform the following simple function as
   part of your application:
  • Given a sparse matrix stored in Compressed Sparse 
  • Column (CSC) format, efficiently compute the 
  • transpose of that matrix. The transpose should 
  • also be stored in CSC format. The input is
  •  provided in three arrays (beg_in[], ind_in[],
  •  and val_in[]), and the result should be returned
  •  in three arrays (beg_out[], ind_out[], and val_out[]).
  •  The matrix dimensions are given by scalars m 
  • (number of rows) and n (number of columns).

The function can be written in any language.

To apply

• To apply for this position, please submit a resume,
•  plus code that implements the routine described above
•  to jobs@gurobi.com.

Post-Doctoral Research Associate Position at University of Edinburgh

Applications are invited for a 2-year Postdoctoral Research Fellow position in Optimization and Stochastics in the School of Mathematics at the University of Edinburgh, United Kingdom. The successful candidate will join the Edinburgh Research Group in Optimization (http://www.maths.ed.ac.uk/ERGO/ ) and participate in the project "Mathematics for Vast Digital Resources". The project is funded by EPSRC, and led by Prof. Jacek Gondzio, Dr. Peter Richtarik and Dr. Burak Buke.

The ideal candidate will be an enthusiastic and creative individual with a Ph.D. in Operational Research, Mathematics or a related field and will have strong publication record. Individuals with a joint background/experience in stochastic programming and large-scale optimization  are particularly encouraged to apply.

The proposed start date for the position is January 1, 2012; however, it can be shifted to a mutually agreed date.  Only complete electronic applications will be considered; a CV, a two-page research plan and at least three references (PDF files preferred) should be submitted. This can be done at http://www.jobs.ed.ac.uk/ using the job reference number 3014768. The closing date for receipt of applications and letters of reference is September 20, 2011.  Interviews of shortlisted candidates will take place in October 2011.

For informal enquiries please contact any of the three project members:

Dr Burak Buke, b.buke@ed.ac.uk

Prof Jacek Gondzio, j.gondzio@ed.ac.uk

Dr Peter Richtarik, peter.richtarik@ed.ac.uk

Couenne Stable Release 0.4 Announced

This is to announce the 0.4 stable version of Couenne. There are a number
of additions and improvements, including:

1) a Feasibility Pump heuristic for non-convex MINLP, developed with Timo
Berthold at the ZIB institute.

2) Orbital Branching for MINLP, developed with Jim Ostrowski and Leo
Liberti.

3) Fixed Point Bound tightening, a bound reduction procedure developed
with Sonia Cafieri, Jon Lee, and Leo Liberti.

4) "semi-auxiliaries", i.e., auxiliary variables defined as y >= f(x) or y
<= f(x) instead of just y = f(x). The purpose is to save on the number of
auxiliaries generated and hence on the size of the LP relaxation.

5) "Two-Implied bound tightening", a new bound reduction procedure
described in http://www.optimization-online.org/DB_FILE/2011/02/2931.pdf

6) various bug fixes.

Release 0.4.0 is a snapshot of the new stable version. The new features
will soon be documented in Couenne's user manual, available at
http://www.coin-or.org/Couenne/couenne-user-manual.pdf