Comments on: Perspectives on integer programming for time-dependent models 2

• Gianpaolo Ghiani ^[1] ; Emanuela Guerriero ^[1]
  1. [1] University of Salento

     University of Salento

     Lecce, Italia

     
• Localización: Top, ISSN-e 1863-8279, ISSN 1134-5764, Vol. 27, Nº. 2, 2019, págs. 178-179
• Idioma: inglés
• DOI: 10.1007/s11750-019-00510-8
• Enlaces
  • Texto completo
• Resumen

  • This paper presents an updated and comprehensive review of discrete optimization techniques for solving time-dependent problems, i.e. decision problems in which activities and resources have to be scheduled over time. Known compact models with continuous variables representing arrival/service/departure times are either (i) nonlinear or (ii) linear with a weak relaxation. This is why practitioners (and a consistent part of the scientific literature) make use of a discretization of time which introduces an approximation. Such extended Integer Programming (IP) formulations rely on binary variables indexed by time (time-indexed models, TI). Their size may be huge, and so they tend to be out of reach for the current IP solver technology.