This document is an attachment of the working paper:
E. Angelelli, C. Archetti and M. Vindigni, The Clustered Orienteering Problem,
WPDEM 2013/7, Working Papers Department of Economics and Management,
University of Brescia.
We document the file format for instances. Files can be downloaded at http://or-brescia.unibs.it/

Instances

Instances are obtained from a set of 57 TSP instances taken from TSPLIB95 library (from 42 to 532 vertices)
available at
http://comopt.ifi.uni-heidelberg.de/software/TSPLIB95
Files are organized as follows.
- COP files: COP problem instances. Each of them is a TSPlib file, with additional data concerning
clusters of customers
- XLSX files: describe the solution found by the algorithms described in the paper
COP files (Instances)
- File name is structured as follows:
<TSP-instance-name>s<#clusters>g<#kind of generation of profits>q<parameter theta
>.cop
- File format follows TSPLIB format with an additional CONSTRAINING_SET section
CONSTRAINING_SET_SECTION:
One line for each cluster of customers:
<cluster_id> <cluster_prize> list:<customer_id>

XLSX files (Results)

Results are reported in:

REPORT-COP-EXACT.xlsx

Two worksheets: COP-CUT, BASIC
800 rows each: (50 graphs ranging from 42 to 318 vertices, s=10,15,20,25, g=1,2, q=2,3)
Worksheets COP-CUT and BASIC have the following columns:
Solver Name either COP-CUT or BASIC
Solver Value solution value found by the solver
Computational Time computing time spent by the solver (in seconds)
Graph graph taken from TSPLIB95
Edge Weight type of edge weight (see TSPLIB95 documentation)
# Vertices number of vertices in the graph
# Clusters number of clusters in the instance
TSP/4 value of parameter theta
Generation kind of generation of profits
Tmax maximum time allowed to the tour
Sum prizes total prize available in the instance
Instance File instances file name
# solcluster number of clusters in the solution
Cluster Indices(+prize) list of comma separated indices of cluster in the solution (+prize of the cluster)
Travel Time duration of the tour in the solution
Tour sequence of customers’ id in the solution tour
Cplex Status final status of CPLEX
LP Relaxation LP relaxation of MIP model
Nodes number of nodes in B&B tree
Best Bound Best proven upper bound
LP-GAP GAP between solution and LP relaxation
UB-GAP GAP between solution and best upper bound

REPORT-COP-TABU.xls

Three worksheets: Basic, Reactive, Multi-Start
924 rows each:
912 rows: (57 graphs ranging from 42 to 532 verticies, s=10,15,20,25, g=1,2, q=2,3)
12 rows: (graph att532, s=50,75,100, g=1,2, q=2,3)
Worksheets Basic, Reactive and Multi-start have the following columns:
Solver Name either COP-TABU Basic or COP-TABU-Multi-start or COP-TABU Reactive
Solver Value solution value found by the solver
Computational Time computing time spent by the solver (in seconds)
Graph graph taken from TSPLIB95
Edge Weight type of edge weight (see TSPLIB95 documentation)
# Vertices number of vertices in the graph
# Clusters number of clusters in the instance
TSP/4 value of parameter theta
Generation kind of generation of profits
Tmax maximum time allowed to the tour
Sum prizes total prize available in the instance
Instance File instances file name
# solcluster number of clusters in the solution
Cluster Indices(+prize) list of comma separated indices of cluster in the solution (+prize of the cluster)
Travel Time duration of the tour in the solution
Tour sequence of customers’ id in the solution tour
IterationFound iteration at which the best solution was found
Sequence list of pairs (iteration, value) updated whenever the algorithm improves on the
incumbent solution.