Picker travel time model for an order picking system with buffers

This article addresses a new class of picker-to-goods order picking systems, in which item retrieval is decoupled from order assembly by a pick-to-buffer technology. This decoupling of item retrieval and order assembly dramatically changes the dynamics of the pick process. Picking to buffers means that the picker is limited by the number of available buffers, rather than by the number of items or orders that may be transported simultaneously, so that order assembly must be carefully synchronized, in real-time, so the picker is not blocked by already filled buffers. This synchronization requirement forces consideration of the stochastic nature of the picking process in designing such systems, in particular, estimating the variance of pick times. This is in sharp contrast with traditional picker-to-goods systems, where estimates of mean pick times are sufficient for evaluating design performance. In this article, models for both expected value and variance of picker travel time are proposed and evaluated. The approximate analytic results are compared with results from detailed simulation and are sufficiently accurate to support design. The real-time synchronization of order assembly with picking is a topic of further research.