Institut[e] f(o|ü)r Informati(cs|k), [Universität] Osnabrück [University]
Here you find the MOSP instances we use in our publications.
Each *.graph file contains one instance of the original set. The files are ASCII encoded. Each node has a natural number as id. The first line is a header line and has the format
<number of nodes> <number of arcs> <number of objectives> <start node id> <target node id>
The rest of the file is organized as an adjacency list of the arcs:
<tail node> <head node> <objective 1 value> <objective 2 value> <objective 3 value> …
The graph is to be interpreted as a directed graph, i.e., an arc from node 1 to node 2 does not imply a connection from node 2 to node 1.
The *.front files are the ASCII encoded nondominated points of the instance. Again, each row contains exactly one vector where the components are space separated.
More information can be found on http://www.mat.uc.pt/~zeluis/INVESTIG/MSPP/DataBase/instanceDataBase.htm.
We mirror the medium and large instances as well as the complete nondominated sets here. However, the instances are encoded in our format above instead the original.
You will find the Copula grid instances here soon.