An extension of a preemptive open-shop scheduling problem is introduce
d. All processing times are integral and in each period i there is a c
ost ci for each task which is processed in that period. Finding a sche
dule with minimum total cost is shown to be NP-hard; some solvable cas
es are discussed; bounds on the cost of an optimum schedule are comput
ed. Finally, a special case is studied, namely where each job has at m
ost three tasks and each processor has to work on at most three tasks.
It is shown to have theoretical complexity equivalent to the general
case.