A project for S. P. Richards Company

The company and its DC operations

P. Viehweg, Senior Vice-President for Logistics of S. P. Richards Company (SPR), visited on 04 February and described the company and some engineering goals at the Philadelphia distribution center. (His presentation is available here in ppt (8MB) or pdf (8 MB) formats.)

Here are highlights from his talk. SPR distributes wholesale office products. Each DC accepts orders until 5PM and ships overnight for next day delivery. The DC's depend on a mostly manual system because it is not clear that capital investment would generate sufficient savings.

Most of the activity in Philadelphia is in zones C and D. Zone C, on the mezzanine (upstairs), holds light-bulk items, such as briefcases. Zone D, downstairs, holds the small parts, such as pens and pencils. (As expected, Zone D holds the most items.) There are also some light-bulk items in Zone D to help complete orders within that zone.

There are four pickers and two checker/packers upstairs in Zone C; and there are eleven pickers and five checker/packers downstairs in Zone D. (During the last hour of work, everyone from Zone C comes downstairs to help complete the work in Zone D.

Each order-picker works within a single zone. He/she pushes a cart and travels a complete circuit through the main aisles of the zone, picking a batch of orders totalling about 50 pick-lines from among about 100 aisles of static shelving. Workers pick directly into final packaging. Most of the time separate boxes are packed in each zone, which means that the customer might receive multiple boxes, some of which are, strictly speaking, unnecessary. Some customers insist on reduced packaging and their orders must be sent from Zone C to Zone D for completion, which greatly increases their flow time through the system.

The project

The Philadelphia DC is reported to be among the best in the SPR system. However they continue to look for ways to improve. They have generated the following list of possible improvements, to which I have added some thoughts to help you get started:

• Increase order-picking efficiency in the C and D zones - to reduce labor costs or increase capacity. This could be done by reorganizing product, especially the small parts on the ground floor, so that it reflects popularity. (Currently product is stored by vendor, and alphabetically by sku ID within vendor.) How much walking is being done now? Can you estimate from the sales history and the warehouse layout? How much might this be reduced if the DC were to move popular products closer towards the main aisles? (From easiest to hardest, consider swapping sections of shelf, shelves, individual sku's.) Should SPR consider using pick-path optimization?
• Reduce the number of multi-zone orders - to reduce packaging and increase customer satisfaction. This could be done on the ground floor by choosing the right light-bulk sku's to put in the 30 inch shelving to complement the small parts. There is the additional question of how much space on the ground floor should be devoted to light-bulk.
• Reduce the amount of empty space in storage - so that more space will be available as the number of sku's increases. This will require accurate dimensional data for all sku's.
• Increase the efficiency of UPS orders.
• Improve operations of full-case versus broken-case picking of paper. Are there sku's for which full-cases should be picked separately from less-than-full-case quantities ("broken-case")? Which generates the most activity? How many orders require both?
• Are there any capital investments that fit the business? Does the conveyor layout make sense?

Data

The company data is copyrighted and proprietary. You may use it for the purposes of this course only. (If you would like to use it for something else, please contact me to discuss.)

You can download all the data below in one monolithic database (zipped mdb format, 30M). Alternatively, you can download the following flat, tab-delimited text files, which have been exported from this database. Finally, I can provide a CD-ROM for anyone without fast internet connection.

• Sales history (zipped, tab-delimited text file, 21.8M). Note that SPR recycles order numbers. The first two digits of the order number represent the Distribution Center number;in this case, all order numbers start with 23 for Philadelphia. The last 6 digits begin at 100000, and when they reach 999999, they start over again at 100000. The order number and the date should make the order unique.
• Item master (zipped, tab-delimited text file, 652.2K)
  • Sku dimensions (zipped, tab-delimited text file, 1.6M)
• Warehouse data
  • Warehouse zones
  • Zone C: Upstairs layout and addresses. (This is the area reserved for light bulkly items, such as briefcases. It is all shelving that is 30 inches deep; and each section of shelf is 42 inches wide.) The odd numbers are on your left as you enter a side aisle and the even numbers are on your right. Sku's are addressed only to the section of shelf (so the order-picker has to hunt within the section). Thus address C08906 means section C, aisle 89, bin 6. Some addresses include the letter A, which identifies the location as a pallet. Generally, though not always, sku's are stored by ID in increasing alphanumeric sequence away from the main aisle.
  • Zone D: Downstairs layout and addresses. (This is the area reserved for small items, such as pens and pencils, plus some light bulk. The small items are stored in shelving that is 18 inches deep; and the light bulk is stored in shelving that is 30 inches deep, on the right end of the warehouse. For both types of shelving each section is 42 inches wide.) Sections labeled A, C, E, etcetera are to your left as you enter a side aisle; and sections B, D, F, etecetera are to your right. The sku's are addressed to the level of shelf, so that D118E5 refers to section D, aisle 118, section E, shelf 5 from the top. Again, the sku's are generally, though not always, stored by ID in increasing alphanumeric sequence away from the main aisle.
  • Bin locations (tab-delimited text file)
  • Two AutoCad drawings: 1, 2.

Be sure you are aware that this, like all data, has flaws

Organization

Order-pickers in Zone D follow the red pick-path and detour as necessary into the cross aisles. Orders start/finish at the head of the conveyor (marked by a red square).

A pick ticket for SPR. The pick ticket number in the upper right hand-corner is the order number in the sales history above. In this example, it is 01-885584. The 01 indicates the Atlanta DC. If this pick ticket were for the Philadelphia DC it would read 23-885584. The left-most column gives the locations to be visited. This example requires picks from zones A, C, and D.

Suggested first steps

• Secure a computer with a database program and some general programming language (recommendation: Java)
• Make sure you understand the meaning of each of the fields in the database
• Generate some initial statistics to check for plausibility. For example: How many sku's are in each zone? How many orders per day? How many lines per order? What are the ten most popular sku's? Ten least popular? What are the most/least popular vendors? How many pick-lines are in each zone?
• In the file of sku's, break the storage addresses apart so that you have a separate field for each component (zone, aisle, section, shelf). This will allow you to locate sku's to any of several degrees of precision.
• Relate addresses to geometry of warehouse layout (so that you will be able to compute the travel distance between two locations along the pick path).
• Build the capability to estimate total travel time as follows. Assume 50 pick-lines per batch. Estimate the travel time for each batch by adding the estimated time to walk the outline of the pick-path plus the deepest travel into and out of each visited aisle.