Scheduling programs with repetitive projects: a comparison of a simulated annealing, a genetic and a pair-wise swap algorithm

The problem of scheduling the one-time production of several units of a single product is examined. To this end, the problem is modelled using two penalty cost structures to account for trade-off between the need to finish each unit by a certain due date and learning-on-the-job effects. In addition, the performances of a simulated annealing procedure, a genetic algorithm and a pair-wise swap heuristic for solving the problem are compared.