TY - JOUR

T1 - The incremental connected facility location problem

AU - Arulselvan, Ashwin

AU - Bley, Andreas

AU - Ljubić, Ivana

N1 - © 2019 Elsevier Ltd. All rights reserved.
Ashwin Arulselvan, Andreas Bley, Ivana Ljubić, The incremental connected facility location problem, Computers & Operations Research, Volume 112, 2019, 104763, https://doi.org/10.1016/j.cor.2019.104763

PY - 2019/12/1

Y1 - 2019/12/1

N2 - We consider the incremental connected facility location problem (incremental ConFL), in which we are given a set of potential facilities, a set of interconnection nodes, a set of customers with demands, and a planning horizon. For each time period, we have to select a set of facilities to open, a set of customers to be served, the assignment of these customers to the open facilities, and a network that connects the open facilities. Once a customer is served, it must remain served in subsequent periods. Furthermore, in each time period the total demand of all customers served must be at least equal to a given minimum coverage requirement for that period. The objective is to minimize the total cost for building the network given by the investment and maintenance costs for the facilities and the network summed up over all time periods. We propose a mixed integer programming approach in which, in each time period, a single period ConFL with coverage restrictions has to be solved. For this latter problem, which is of particular interest in itself, new families of valid inequalities are proposed: these are set union knapsack cover (SUKC) inequalities, which are further enhanced by lifting and/or combined with cut-set inequalities, which are primarily used to ensure connectivity requirements. Details of an efficient branch-and-cut implementation are presented and computational results on a benchmark set of large instances are given, including examples of telecommunication networks in Germany

AB - We consider the incremental connected facility location problem (incremental ConFL), in which we are given a set of potential facilities, a set of interconnection nodes, a set of customers with demands, and a planning horizon. For each time period, we have to select a set of facilities to open, a set of customers to be served, the assignment of these customers to the open facilities, and a network that connects the open facilities. Once a customer is served, it must remain served in subsequent periods. Furthermore, in each time period the total demand of all customers served must be at least equal to a given minimum coverage requirement for that period. The objective is to minimize the total cost for building the network given by the investment and maintenance costs for the facilities and the network summed up over all time periods. We propose a mixed integer programming approach in which, in each time period, a single period ConFL with coverage restrictions has to be solved. For this latter problem, which is of particular interest in itself, new families of valid inequalities are proposed: these are set union knapsack cover (SUKC) inequalities, which are further enhanced by lifting and/or combined with cut-set inequalities, which are primarily used to ensure connectivity requirements. Details of an efficient branch-and-cut implementation are presented and computational results on a benchmark set of large instances are given, including examples of telecommunication networks in Germany

KW - mixed integer programming

KW - incremental network design

KW - multi-period network design

KW - branch-and-cut

KW - facility location

UR - https://www.sciencedirect.com/journal/computers-and-operations-research

U2 - 10.1016/j.cor.2019.104763

DO - 10.1016/j.cor.2019.104763

M3 - Article

SN - 0305-0548

VL - 112

SP - 1

EP - 17

JO - Computers & Operations Research

JF - Computers & Operations Research

M1 - 104763

ER -