DBSLAP Data Set

Published: 16 February 2019| Version 2 | DOI: 10.17632/fybkpf9f6n.2
Contributor:
Cihan Tugrul Cicek

Description

This repository contains a collection of benchmark instances for the Drone Base Station Location-Allocation Problem (DBSLAP), which is an extension to covering location problems, introduced in the paper titled "The location-allocation of drone base stations". In the DBSLAP, there are m Macro-Base Station (MBS) that provides resource to p Drone Base Stations (DBS) that serve to n users. There exists a nonnegative activation cost, v, for each DBS and a nonnegative variable allocation cost, z, for each MBS. Each MBS has a resource capacity, Phi, that is independent of which DBS is supported. Each user yields a nonnegative unit revenue, c, when served. The objective is to serve all users with at least a minimum service requirement to maximize the total profit given the coordinates of users and MBSs. This problem has been attracting significant interest as the launch of 5th. generation wireless networks is expected to allow telecommunication companies to establish heterogenous network structures, which do not only include terrestrial base stations but also aerial base stations. If you find any errors or like to comment, or if you have additional benchmarks or code material, please write an e-mail to cihantugrulcicek[at]gmail.com.

Files