Article ID: | iaor2017213 |
Volume: | 68 |
Issue: | 1 |
Start Page Number: | 62 |
End Page Number: | 73 |
Publication Date: | Jan 2017 |
Journal: | J Oper Res Soc |
Authors: | Hertz Alain, Marcotte Odile, Mdimagh Asma, Carreau Michel, Welt Franois |
Keywords: | combinatorial optimization, location, networks, allocation: resources, programming: multiple criteria, programming: integer, programming: quadratic, simulation |
In this article we consider a real‐world problem submitted to us by the Hatch company. This problem consists of designing a collection network for a wind farm, assuming that the locations of the turbines and the potential cables are known, several cable types are available, and the cost of the energy that dissipates through the cables is known. We propose a mixed integer quadratic programme to model the network design problem and then linearize the quadratic programme because the latter is too difficult to solve using a standard mathematical programming software. We describe several classes of inequalities that strengthen the resulting mixed integer linear programme. Finally we use real‐world data supplied by Hatch to carry out computational experiments with several versions of our model.