A Combined Data Mining Based-Bi Clustering and Order Preserved Sub-Matrices Algorithm for Set Covering Problem

This study evaluates a Set Covering Problem (SCP), an extension of the demand covering problem, with several potential applications. The original demand covering problem objective includes the selection of proper locations for a number of available facilities to cover the required demand. The SCP tries to minimize location cost satisfying a specified level of coverage. The SCP problems answer many location problems, e.g., the emergency services sector with alternative facilities that will cover the unavailability of the primary facility or recommender systems where it is desired to fulfill the demand by several available choices. We present a biclustering method to construct biclusters from the distance matrix where a bicluster depicts a subset of demand centers covered by a subset of facilities. According to experiments performed in this study, it is concluded that the proposed method provides high-quality solutions compared with an optimal solution attained from GAMS. Also, for larger problem instances, the proposed method provided solutions with higher quality than GAMS software when the computational time is limited to 1 Hour.