One of the challenges facing e-commerce fulfillment is the rigidity of the wave management system that often dictates a fixed batch size that must be processed before the next wave of orders is released. As a result, operation managers usually push multiple waves, which lead to accumulation at bottleneck resources, and poor synchronization at sortation engines. We demonstrate the power of operating an order fulfillment engine using the “pull” paradigm. We hypothesize that a pull-driven order fulfillment can meet the same throughput requirements achieved by the push-driven wave-based order fulfillment, but with higher speed, higher service levels, more level resource utilization, and smaller hardware investment cost. To that end, we develop analytical throughput models of a zone picking operation, followed by a sortation operation to separate and consolidate the orders before packing. We develop an algorithm for determining the picking batch size based on the available capacity at each process, and an algorithm for determining the maximum order threshold to maintain in the system before authorizing release of orders towards picking. We use a discrete-event simulation model to validate our approach and measure the performance differences between pull vs. push control.
Progress in Material Handling Research
Nazzal, Dima; Shanmugasundaram, Shankar; Kim, Sung; and Meller, Russell D., "Pull-driven Order Fulfillment in Distribution Centers" (2018). 15th IMHRC Proceedings (Savannah, Georgia. USA – 2018). 24.