Distribution Center Slotting and Optimization Using Simulation

Hosni Adra

Director of Simulation and Process Improvement
CreateASoft, Inc

Simulation and optimization have been used for many years to improve internal operations like receiving and shipping in warehouses and distribution centers. Most of these simulation environments are built on distributions and potential scenarios, rather than actual warehouse flows. This case study presents an improved method to optimize the warehouse using an actual outbound dataset that takes into consideration seasonality, SKU slotting, rack types, and pick path. Moreover, the generated simulation model is used for slotting analysis and optimization.

The analyzed distribution center consists of three main areas: a freezer, a cooler, and a general rack pick area. All orders that are picked will be delivered to a limited space staging area before they are put on trucks for delivery. Received pallets are consolidated and sent to their appropriate locations based on type. Shelves are replenished from top to bottom when needed. Picking is always from the bottom two shelves, also known as the strike zone.

All receiving occurs before 10AM. Pallets are unloaded, consolidated when needed, then stored in top shelves. All receiving activities are completed by 11 AM. During the storage/receiving process, all bottom shelves are replenished to full capacity. Picking begins at 11 AM and is split into two main components: Parcels and LTL orders. Parcels are small orders that normally fit in a few boxes and are shipped through overnight delivery. LTL orders are palletized for truck delivery.

The goal of the simulation was to resolve a number of issues in the operation:

• Receiving issues based on the ten docks currently available. Define a method where receiving consistently completes by 11AM. The current situation is prone to truck delays, consolidation, movement to storage, and the size of the staging area.
• Improving the overall order release process to minimize the outbound staging and optimize the pick path. The main goal for pick path optimization is in reducing the travel distance, reducing the empty resource travel, and maximizing the grab factor.
• Re-slotting the warehouse as needed in order to achieve the first 2 issues.

With Simcad Pro, an interactive on-the-fly simulator, a full representation of the warehouse is built that includes the following:

• identifying each rack location within the model, while allowing multiple SKUs to be in a single rack space;
• defining the racking based on the physical characteristics in volume and functionality;
• defining the aisle of travel to allow pickers traverse the warehouse and perform the picking cycle; and
• expanding the docks and staging area to perform both inbound and outbound staging.

In order to simplify the data transition from the WMS (Warehouse Management System) to the simulation model, all existing warehouse rack and SKU naming conventions where used. From a modeling perspective, the modeler used the existing rack and path building wizards in Simcad Pro to create the warehouse structure. Moreover, an existing CAD layout of the facility was used as a background layer and to properly compute the travel distances.

The model validation process requires the run of the model for a full year in order to account for seasonality. Hence, the model was initially validated using a single week of operation, amounting to about 750,000 transactions. The data was directly loaded from a WMS system. Validation consisted of two phases:

• Running the model based on the WMS generated pick sequence and storage moves. The model used built-in constraints to drive the WMS simulation. The resulting validation shows a model accuracy of 99.91% as compared to the real-time performance of the analyzed week. The process was repeated for four weeks, accounting for seasonality change, and the model’s accuracy was within 0.02% of the initial week run accuracy.
• Starting the model with basic outbound and inbound datasets. The imported dataset is used to create the orders, picking, consolidation, and storage based on the defined constraints. This type of data generation helped in expanding the model in order to quickly perform increased capacity analysis and re-slotting.

Before proceeding to the warehouse analysis phase, a detailed spaghetti diagram, congestion analysis, and heat map were generated through the Simcad Pro interface. All maps and diagrams were generated per pick zone and operation type.

The following are the results based on the analysis performed:

• The previous slotting implementation did not take into consideration the grab factor and congestion generated by the pick path. Using Simcad Pro, two paths slotting optimization, and a re-slotting of each zone resulted in a 7%, 9%, and 6.5% efficiency increase in the pick path for the freezer, cooler, and floor pick areas respectively.
• Replenishment cycles were reduced by 35% throughout the warehouse. This optimization did not require any additional racking.
• The inbound staging area was reorganized based on truck arrival and the type or location of the product. Specific rules regarding when to dock, the number of personnel in the consolidation area, and put-away personnel was defined.
• Order release to the freezer, cooler, and floor pick area were modified to correspond to the expected pick time of each order. Order release was performed based on the each order pick duration.

All modifications resulted in a 22.3% reduction in congestion in the outbound area and decreased truck dock-to-ship time by 15 minute on average.

Adra, H. 2019 Success With Simulation – A definitive guide to process improvement success using simulation for healthcare, manufacturing and warehousing. 1st ed.
Aurora: Adra
Simcad Pro – Interactive, On-The-Fly simulation software.

PDF DOWNLOAD