<p>In scheduling problems with generalized due-dates, the due-dates are position-dependent (and not job-dependent as in classical scheduling). In this paper, we study scheduling problems on parallel machines, and the underlying assumption is that the generalized due-dates are machine-dependent. The following scheduling measures are considered: total tardiness, maximum tardiness, number of tardy jobs, and total late work. We show that all the problems are NP-hard even if all generalized due-dates are identical. We complement this hardness result by showing that all problems are solvable in pseudo-polynomial time and that minimizing total late work is fixed parametrized tractable with respect to the number of different generalized due-dates and processing times in the instance. We also tested the pseudo-polynomial time algorithms, showing they can easily solve instances containing up to 200 jobs.</p>