A polling model for an order-picking workstation

We investigate the performance of parts-to-picker warehouses with remotely located order-picking workstations. More specifically, we deduce mathematical expressions for the order flow times, which are defined as the time between release and completion of a customer order in the warehouse. To this end, we model an order-picking workstation as a polling system so that order flow times correspond to cycle times in the polling system. The order-picking policy of each order picker handling one order at a time to avoid picking errors is captured by a specific service discipline, which, to the best of our knowledge, has not been studied before. As this service discipline does not satisfy the so-called branching property, we establish stochastic bounds for the order flow times. These bounds are shown to be adequate and aid in setting targets for the throughput of the storage area, which, in turn, is the input rate of the workstations. As research on parts-to-picker warehouses has traditionally focused on optimizing operations in the storage area, we believe our results are complementary to those established in literature.