Bin Packing with Linear Usage Costs --- An Application to Energy Management in Data Centres

摘要: EnergeTIC is a recent industrial research project carried out in Grenoble on optimizing energy consumption data-centres. The efficient management of data-centre involves minimizing costs while ensuring service quality. We study the problem formulation proposed by EnergeTIC. First, we focus key sub-problem: bin packing with linear associated use bins. lower bounds based Linear Programming and extend global constraint cost information. Second, present column generation model for computing bound original where pricing essentially cost-aware side constraints. Third, show that benchmark provided so far can be solved to near optimality using large neighborhood search. The authors acknowledge their partners (Bull, Schneider Electric, Business & Decision UXP) as well several public institutions (G2Elab, G-SCOP LIG). from UCC are supported Science Foundation Ireland Grant No. 10/IN.1/I3032.