N4M wrote:
> C++ is currently a dominant programming lanaguage to solve
> Combinatorial Optimization problems. Will other languages be able to
> compete with C++ in this field?
IMO, none of the other popular languages (C, Java, C#, Fortran,
Perl, Python, ...) is well suited to implement non-trivial
algorithms. It can be done in these languages but is unnecessary
hard. Whether this will change in the future depends on the
languages. I think support for generic programming is required
for effective support of non-trivial algorithms.
> Could you suggest some C++ libraries for Combinatorial Optimization?
The Boost Graph Library (BGL) has algorithms for several
combinatorial problems. Of course, those implemented can be
solved in polynomial time. There is no general solver for
combinatorial optimization in this library (like an implementation
of the Simplex algorithm).
> It is probably that a library will address a certain problem, e.g.
> Max-Cut, TSP etc., that would already be fine.
It is unlikely that anybody implements something addressing these
problems: the problems you mention are NP-hard and thus only
solvable for very small problem instances or in cases where
additional structure information is available and can be exploited
(e.g. TSP can be solved if certain restrictions on the distances
are imposed - if I remember correctly).
--
<private.php?do=newpm&u=> <http://www.dietmar-kuehl.de/>
<http://www.contendix.com> - Software Development & Consulting
Similar Threads
