Cyclic schedules for pipeline assembly processes

In this work we deal with a resource constrained project scheduling problem. The considered activities represent tasks in a manufacturing assembly process, in which multiple copies of a certain product have to be created. For each copy of the product, the operations must be processed on dedicated machines and must satisfy a given set of precedence constraints. The objective is to minimize the overall makespan. When the number of copies becomes large, we are interested in designing periodic (or cyclic) schedules, where we want to maximize throughput, i.e., number of copies produced in a time unit, while keeping product flow time low. We present a heuristic algorithm based on a branch and bound scheme for the case of a pipeline assembly process.