This paper deals with the permutation flowshop scheduling problem in which there are sequence dependent setup times on each machine, commonly known as the SDST flowshop. The optimisation criterion considered is the minimisation of the makespan or Cmax. Genetic algorithms have been successfully applied to regular flowshops before, and the objective of this paper is to assess their effectiveness in a more realistic and complex environment. We present two advanced genetic algorithms as well as several adaptations of existing advanced metaheuristics that have shown superior performance when applied to regular flowshops. We show a calibration of the genetic algorithm's parameters and operators by means of a Design of Experiments approach. For evaluating the proposed algorithms, we have coded several, if not all, known SDST flowshop specific algorithms. All methods are tested against an augmented benchmark based on the instances of Taillard. The results show a clear superiority of the algorithms proposed, especially for the genetic algorithms, regardless of instance type and size.
