Schedules for pick-and-place robots

Abstract The efficiency of pick-and-place robots that put parts into boxes relies on efficient schedules, which determine the order in which parts are picked. The scheduling problem can be formulated as an integer linear program (ILP). However, the resulting ILP cannot be solved for realistic scenarios. Therefore, approximate solution techniques based on local search heuristics and simplified ILPs are introduced. It will be shown that ILP-based solutions outperform search heuristics in terms of solution times and solution quality. Additionally, we consider an extended version of the pick-and-place robot with two arms. It is shown that the resulting scheduling problem can be handled similarly to the problem for the robot with one arm, but the number of variables in the resulting ILP increases from $$O(P^3)$$ O ( P 3 ) for one arm to $$O(P^6)$$ O ( P 6 ) for two arms, where P is the number of parts to be picked. Consequently, decomposition-based approaches are introduced that reduce the scheduling problem for two arms to several problems with one arm.