Description

The algorithm described in this paper was tested on five data sets, as described in Section 5 of the paper.  The data sets are each contained in the Excel files below.  The geographical data sets (49, 88, and 150 nodes) are adapted from Daskin (1995).

In each data set, the first row indicates the number of nodes (N), the objective function weight (alpha), and the failure probability (q).  Each successive row contains data for one node, including: index, demand, emergency cost, failable (0/1), fixed cost, latitude, and longitude. 

The columns marked "50% F" gives an alternate set of values for the "failable?" column in which approximately 50% of the facilities are failable.  These values were used for the tests described in Section 5.4.

You may download the data sets using the links below.  I would appreciate your filling out and submitting the following form before downloading; however, if you wish to download the data sets anonymously, simply skip the form and click the link below. (When you receive the confirmation page, navigate back to this page and click on the links below to download the data sets.)

Name:
E-mail Address:
University/Company/Affiliation:
Department:
City and State/Provice:
Country:
Comments:

Click a file below to download it, or click here to download a Zip archive containing all five data sets.

• data49UFLP.xls: 49-node great-circle data set (48 continental U.S. state capitals plus Washington, DC; 1990 census data)
  
• data50EucUFLP.xls: 50-node Euclidean data set (random)
  
• data88UFLP.xls: 88-node great-circle data set (49-node set plus 50 largest U.S. cities minus duplicates; 1990 census data)
  
• data100EucUFLP.xls: 100-node Euclidean data set (random)
  
• data150UFLP.xls: 150-node great-circle data set (150 largest U.S. cities; 1990 census data)