General variable neighborhood search for the parallel machine scheduling problem with two common servers (Q6063516)

The authors consider the makespan minimization problem on parallel machines with loading and unloading servers. For any job there is no delay between loading, processing, and unloading operations. It is also supposed that the processing time of any job is strictly less than the total duration of loading, processing, and unloading operations of any other job. For the considered problem the authors develop a mixed-integer-programming model, four lower bounds, and two greedy heuristics. They also propose a general variable neighborhood search meta-heuristic. Numerical experiments are provided for examples with the number of jobs not exceeding 250, the number of machines not exceeding 5. Processing, loading, and unloading times are generated from the uniform distribution between 10 and 100, 5 and 25, 5 and 25 correspondingly.