Scheduling of inventory releasing jobs to satisfy time-varying demand

This paper studies a new class of single-machine scheduling problems, that are faced by Just-in-Time-suppliers satisfying a given demand. In these models the processing of jobs leads to a release of a predefined number of product units into inventory. Consumption is triggered by predetermined time-varying, and product-specific demand requests. While all demands have to be fulfilled, the objective is to minimize the resulting product inventory. We investigate different subproblems of this general setting with regard to their computational complexity. For more restricted problem versions (equal processing times and equal number of released products) strongly polynomial time algorithms are presented. In contrast to this, NP-hardness in the strong sense is proven for more general problem versions (varying processing times or varying number of released products). Moreover, for the most general version, even finding a feasible solution is shown to be strongly NP-hard.