Large neighborhood for a parallel scheduling problem with earliness-tardiness penalties and a common due date

International audienceThis paper addresses the parallel machine scheduling problem where jobs have different earliness-tardiness penalties and a common due date. Since the problem is conjectured NP-hard in the strong sense, we propose a large neighborhood search that can be explored by a pseudo polynomial dynamic program algorithm. Computational experiments are presented and compared to a lower bound in order to verify that their gap is reasonable. They are also compared to simple local search based on moves and swaps between jobs