Local Search: Methodology and Industrial Applications
Frédéric Gardi, Thierry Benoist, Julien Darlay, Bertrand Estellon, Romain Megel · 2014
Many papers describing methodologies or good practices to engineer local search heuristics have been published. This chapter presents methodology for tackling combinatorial or even mixed-variable optimization problems by local search. The methodology is designed to industrialize the engineering of local search heuristics, especially to solve large-scale combinatorial problems encountered in real-world situations. Having exposed the methodology, the chapter illustrates two challenging industrial applications: car sequencing for painting and assembly lines (combinatorial optimization); vehicle and inventory routing (mixed-variable optimization). The chapter provides a simple recipe to help practitioners deliver quality, fast, reliable and robust solutions to their clients, while lowering their development and maintenance costs as well as the risks on their optimization projects.