dPDOP - Solver
d-place dicourse ordering problem solver
For all questions concerning the software, please contact
Ernst Althaus.
You can download this paper for a description
of the problem.
The copyright of this software belongs partly to the
Max-Planck-Society.
We provide our (rudamentary) implementations of our 2PDOP and
3PDOP solvers. To compile the source file, you need an installation
of the LEDA-Library
and the
SCIL-Software package
(which requires some further libraries).
We also provide executables for different systems. For every system, we
have executables for 2PDOP and 3PDOP based on the commercial
CPLEX solver and based on SCIP, which is
not officially released yet and will be free for all academic purposes.
If you need an executable for another system, please send me an email.
Both programs have one parameter with the file containing the instance.
The file format is described below.
The file format
First of all: all empty lines or lines starting with a white-space or a
'#' are ignored.
Notice that the numbering is done in C++-style, i.e. we
start with the number 0.
Other than that, the file of an dPDOP-Instnace has to looks as follows:
(int: number i of discourses)
(string: discourse 0)
(string: discourse 1)
...
(string: discourse i-1)
(The Cost Function)
The cost function for 3PDOP-Instances is encoded as follows:
for every value
$c_I(u_i, u_j)$ of the initial cost function, we add a line
-1 (int: i) (int: j) (double: c_I(u_i, u_j))
and for every value
$c_T(u_i, u_j, u_k)$ of the transition cost function, we add a line
of the format
(int: i) (int: j) (int: k) (double c(u_i, u_j, u_k))
The cost function for 2PDOP-Instances is encodes similary.
For exaples, look at the 3-PDOP-example
and the 2-PDOP-example